feat(flights): cumulative fuel burned + CO2 emitted per flight

Pre-fix the emissions tooltip only showed the per-hour *rate* — what most
users actually want is the cumulative *amount* burned. This adds running
totals computed by multiplying the model-based rate by the elapsed
observation time since we first saw the airframe.

New module ``flight_observations.py``:
* Tracks first_seen_at + last_seen_at per icao24 hex.
* Re-opens a fresh session when an aircraft is unseen for > 15 min
  (treated as a new flight — landed and took off, or transited a dead
  zone). Prevents the cumulative counter from resetting mid-flight if
  the trail-rendering cache prunes the trail.
* Clamps elapsed time to 24h max so clock skew can't produce comically
  large numbers.
* Pruned every 5 min via a new scheduler job (mirrors ais_prune cadence).

flights.py + military.py emission enrichment now also attaches:
* observed_seconds — how long we've been tracking this airframe.
* fuel_gallons_burned — rate * elapsed_h.
* co2_kg_emitted — rate * elapsed_h.

The existing per-hour rate fields stay in the dict for backward compat
and are shown as small secondary context in the tooltip.

Frontend EmissionsEstimateBlock (NewsFeed.tsx) now prominently shows
the cumulative totals with the rate as smaller context underneath plus
"Observed in flight for Xh Ym". When observed_seconds is 0 (first refresh)
it renders "Just observed · totals will appear on next refresh" instead
of a misleading "0 gal".

12 backend tests cover record/accumulate/reset, the 24h clamp, prune,
case-insensitive key normalization, and end-to-end emission integration
in _classify_and_publish.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

backend/.env.example

@@ -11,6 +11,13 @@ AIS_API_KEY=              # https://aisstream.io/ — free tier WebSocket key
 
 # ── Optional ───────────────────────────────────────────────────
 
+# AISHub REST fallback. Used when stream.aisstream.io is unreachable
+# (e.g. their cert expires or server goes offline). Free tier requires
+# registration at https://www.aishub.net/api. Poll cadence defaults to
+# 20 min to stay courteous; tunable via AISHUB_POLL_INTERVAL_MINUTES.
+# AISHUB_USERNAME=
+# AISHUB_POLL_INTERVAL_MINUTES=20
+
 # Override allowed CORS origins (comma-separated). Defaults to localhost + LAN auto-detect.
 # CORS_ORIGINS=http://192.168.1.50:3000,https://my-domain.com
 

backend/main.py

@@ -1417,6 +1417,29 @@ def _peer_sync_response(peer_url: str, body: dict[str, Any]) -> dict[str, Any]:
         proxy = f"socks5h://127.0.0.1:{socks_port}"
         kwargs["proxies"] = {"http": proxy, "https": proxy}
     response = _requests.post(f"{normalized}/api/mesh/infonet/sync", **kwargs)
+    # HTTP 429 must be surfaced as a typed exception carrying the
+    # Retry-After value, so finish_sync can honor it and stop hammering
+    # the upstream. Pre-fix this path just stringified the status into
+    # a ValueError, which finish_sync then ignored — keeping the
+    # upstream's rate-limit bucket full indefinitely.
+    if response.status_code == 429:
+        from services.mesh.mesh_infonet_sync_support import (
+            PeerSyncRateLimited,
+            parse_retry_after_header,
+        )
+
+        retry_after_s = parse_retry_after_header(
+            response.headers.get("Retry-After", "") or "",
+        )
+        try:
+            body_text = response.text[:200]
+        except Exception:
+            body_text = ""
+        raise PeerSyncRateLimited(
+            f"HTTP 429 from {normalized} (retry_after={retry_after_s}s): {body_text}",
+            retry_after_s=retry_after_s,
+            status=429,
+        )
     try:
         payload = response.json()
     except Exception as exc:
@@ -1462,8 +1485,23 @@ def _hydrate_gate_store_from_chain(events: list[dict]) -> int:
     return count
 
 
-def _sync_from_peer(peer_url: str, *, page_limit: int = 100, max_rounds: int = 5) -> tuple[bool, str, bool]:
+def _sync_from_peer(
+    peer_url: str,
+    *,
+    page_limit: int = 100,
+    max_rounds: int = 5,
+) -> tuple[bool, str, bool, int]:
+    """Sync the local Infonet chain against ``peer_url``.
+
+    Returns ``(ok, error, forked, retry_after_s)``. The fourth tuple
+    element is non-zero only when the peer responded with HTTP 429
+    and supplied a parseable ``Retry-After`` header — see the typed
+    ``PeerSyncRateLimited`` exception in mesh_infonet_sync_support.py.
+    Callers should pass that value to ``finish_sync(retry_after_s=...)``
+    so the next attempt actually waits.
+    """
     from services.mesh.mesh_hashchain import infonet
+    from services.mesh.mesh_infonet_sync_support import PeerSyncRateLimited
 
     rounds = 0
     while rounds < max_rounds:
@@ -1472,7 +1510,11 @@ def _sync_from_peer(peer_url: str, *, page_limit: int = 100, max_rounds: int = 5
             "locator": infonet.get_locator(),
             "limit": page_limit,
         }
-        payload = _peer_sync_response(peer_url, body)
+        try:
+            payload = _peer_sync_response(peer_url, body)
+        except PeerSyncRateLimited as exc:
+            # Bubble up the retry-after so finish_sync can honor it.
+            return False, str(exc), False, exc.retry_after_s
         if bool(payload.get("forked")):
             # Auto-recover small local forks: if the local chain is tiny
             # (< 20 events) and the remote has a longer chain, reset local
@@ -1488,23 +1530,23 @@ def _sync_from_peer(peer_url: str, *, page_limit: int = 100, max_rounds: int = 5
                 )
                 infonet.reset_chain()
                 continue  # retry sync with clean genesis locator
-            return False, "fork detected", True
+            return False, "fork detected", True, 0
         events = payload.get("events", [])
         if not isinstance(events, list):
-            return False, "peer sync events must be a list", False
+            return False, "peer sync events must be a list", False, 0
         if not events:
-            return True, "", False
+            return True, "", False, 0
         result = infonet.ingest_events(events)
         _hydrate_gate_store_from_chain(events)
         rejected = list(result.get("rejected", []) or [])
         if rejected:
-            return False, f"sync ingest rejected {len(rejected)} event(s)", False
+            return False, f"sync ingest rejected {len(rejected)} event(s)", False, 0
         if int(result.get("accepted", 0) or 0) == 0 and int(result.get("duplicates", 0) or 0) >= len(events):
-            return True, "", False
+            return True, "", False, 0
         if len(events) < page_limit:
-            return True, "", False
+            return True, "", False, 0
         rounds += 1
-    return True, "", False
+    return True, "", False, 0
 
 
 def _run_public_sync_cycle() -> SyncWorkerState:
@@ -1567,11 +1609,12 @@ def _run_public_sync_cycle() -> SyncWorkerState:
         with _NODE_RUNTIME_LOCK:
             set_sync_state(started)
         try:
-            ok, error, forked = _sync_from_peer(record.peer_url)
+            ok, error, forked, retry_after_s = _sync_from_peer(record.peer_url)
         except Exception as exc:
             ok = False
             error = str(exc or type(exc).__name__)
             forked = False
+            retry_after_s = 0
         if ok:
             store.mark_seen(record.peer_url, "sync", now=time.time())
             store.mark_sync_success(record.peer_url, now=time.time())
@@ -1618,6 +1661,12 @@ def _run_public_sync_cycle() -> SyncWorkerState:
             now=time.time(),
             interval_s=int(get_settings().MESH_SYNC_INTERVAL_S or 300),
             failure_backoff_s=failure_backoff_s,
+            # 429 retry-storm fix: when the peer returned HTTP 429 with
+            # a Retry-After header, finish_sync uses max(exponential,
+            # retry_after) for next_sync_due_at — so we actually wait
+            # the time the upstream asked for instead of hammering
+            # every 60s and keeping its rate-limit bucket full forever.
+            retry_after_s=retry_after_s,
         )
         with _NODE_RUNTIME_LOCK:
             set_sync_state(updated)

backend/routers/ai_intel.py

@@ -2521,45 +2521,85 @@ async def api_capabilities(request: Request):
 # OpenClaw Connection Management (local-operator only — NOT via HMAC)
 # These endpoints manage the HMAC secret itself, so they MUST require
 # local operator access to prevent privilege escalation.
+#
+# Issue #302 (tg12): pre-fix, GET /api/ai/connect-info had two problems:
+#
+#   1. ``?reveal=true`` made the full secret travel through every operator
+#      page-load that opened the Connect modal. Even gated to
+#      ``require_local_operator``, that put the secret into browser
+#      history, dev-tools network panels, browser disk caches, HAR
+#      exports, and screen captures. Every time the modal opened.
+#
+#   2. The same GET endpoint auto-bootstrapped (generated + persisted)
+#      the secret on first read. Side effects on a GET are a footgun:
+#      browser prefetchers, mirror tools, and casual curl-from-history
+#      would all silently mint+persist a fresh secret. (Gated, but
+#      still surprising — and noisy in the audit log.)
+#
+# Resolution:
+#
+#   GET  /api/ai/connect-info             — always returns the MASKED
+#                                            secret. No ?reveal param.
+#                                            No auto-bootstrap; if the
+#                                            secret is missing,
+#                                            ``hmac_secret_set: false``
+#                                            tells the frontend to call
+#                                            /bootstrap.
+#
+#   POST /api/ai/connect-info/bootstrap   — NEW. Generates + persists the
+#                                            secret if missing. Idempotent.
+#                                            Returns metadata only, never
+#                                            the full secret.
+#
+#   POST /api/ai/connect-info/reveal      — NEW. Returns the full secret in
+#                                            the body with strict
+#                                            ``Cache-Control: no-store,
+#                                            no-cache, must-revalidate``
+#                                            + ``Pragma: no-cache`` so
+#                                            it does not land in browser
+#                                            caches. POST means it does
+#                                            not land in URL history.
+#
+#   POST /api/ai/connect-info/regenerate  — keeps existing one-time-reveal
+#                                            behavior (regenerate IS a
+#                                            deliberate destructive action
+#                                            the operator triggered, so
+#                                            displaying the new secret
+#                                            once is the only path that
+#                                            makes the operation useful).
+#                                            Same no-store headers added.
 # ---------------------------------------------------------------------------
 
-@router.get("/api/ai/connect-info", dependencies=[Depends(require_local_operator)])
-@limiter.limit("30/minute")
-async def get_connect_info(request: Request, reveal: bool = False):
-    """Return connection details for the OpenClaw Connect modal.
+# Cache-Control headers that should accompany every response carrying the
+# full HMAC secret. Reused across the reveal + regenerate endpoints so a
+# future refactor that splits or renames them can't forget the headers.
+_NO_STORE_HEADERS = {
+    "Cache-Control": "no-store, no-cache, must-revalidate, private",
+    "Pragma": "no-cache",
+    "Expires": "0",
+}
 
-    The HMAC secret is masked by default. Pass ?reveal=true to see the full key.
-    Private keys are NEVER returned.
+
+def _mask_hmac_secret(secret: str) -> str:
+    """Return a fingerprint-style mask (first6 + bullets + last4) suitable
+    for display in the UI before the operator clicks Reveal."""
+    if not secret:
+        return ""
+    if len(secret) > 10:
+        return secret[:6] + "••••••••" + secret[-4:]
+    return "••••••••"
+
+
+def _connect_info_metadata(settings) -> dict:
+    """Return everything the Connect modal needs EXCEPT the secret itself.
+
+    Shared between GET /api/ai/connect-info (where the full secret is
+    masked) and POST /api/ai/connect-info/bootstrap (where the operator
+    just generated a secret but we don't return it inline — they have to
+    call /reveal to see it).
     """
-    import os
-    import secrets
-    from services.config import get_settings
-
-    settings = get_settings()
-    hmac_secret = str(settings.OPENCLAW_HMAC_SECRET or "").strip()
     access_tier = str(settings.OPENCLAW_ACCESS_TIER or "restricted").strip().lower()
-
-    # Auto-generate if not set
-    if not hmac_secret:
-        hmac_secret = secrets.token_hex(24)  # 48 chars
-        _write_env_value("OPENCLAW_HMAC_SECRET", hmac_secret)
-        # Clear settings cache so next read picks up the new value
-        get_settings.cache_clear()
-
-    masked = hmac_secret[:6] + "••••••••" + hmac_secret[-4:] if len(hmac_secret) > 10 else "••••••••"
-
     return {
-        "ok": True,
-        "hmac_secret": hmac_secret if reveal else masked,
-        "hmac_secret_set": bool(hmac_secret),
-        "bootstrap_behavior": {
-            "auto_generates_when_missing": True,
-            "auto_generated_this_call": not bool(settings.OPENCLAW_HMAC_SECRET or ""),
-            "notes": [
-                "If no HMAC secret exists yet, this endpoint bootstraps one and persists it to .env.",
-                "Regenerating the HMAC secret revokes all existing direct-mode OpenClaw callers at once.",
-            ],
-        },
         "access_tier": access_tier,
         "trust_model": {
             "remote_http_principal": "holder_of_openclaw_hmac_secret",
@@ -2613,24 +2653,138 @@ async def get_connect_info(request: Request, reveal: bool = False):
     }
 
 
-@router.post("/api/ai/connect-info/regenerate", dependencies=[Depends(require_local_operator)])
-@limiter.limit("5/minute")
-async def regenerate_hmac_secret(request: Request):
-    """Generate a new HMAC secret. Old secret immediately stops working."""
+@router.get("/api/ai/connect-info", dependencies=[Depends(require_local_operator)])
+@limiter.limit("30/minute")
+async def get_connect_info(request: Request):
+    """Return connection details for the OpenClaw Connect modal.
+
+    The HMAC secret is always returned as a fingerprint mask
+    (``first6 + bullets + last4``); the full value is only ever served by
+    ``POST /api/ai/connect-info/reveal`` (see #302). When the secret has
+    not been bootstrapped yet, ``hmac_secret_set`` is false and the
+    frontend should call ``POST /api/ai/connect-info/bootstrap``.
+
+    Private keys are NEVER returned.
+    """
+    from services.config import get_settings
+
+    settings = get_settings()
+    hmac_secret = str(settings.OPENCLAW_HMAC_SECRET or "").strip()
+
+    return {
+        "ok": True,
+        "masked_hmac_secret": _mask_hmac_secret(hmac_secret),
+        "hmac_secret_set": bool(hmac_secret),
+        "bootstrap_behavior": {
+            "auto_generates_when_missing": False,
+            "notes": [
+                "Call POST /api/ai/connect-info/bootstrap to mint a secret on first use.",
+                "Call POST /api/ai/connect-info/reveal to see the full secret (no-store).",
+                "Regenerating the HMAC secret revokes all existing direct-mode OpenClaw callers at once.",
+            ],
+        },
+        **_connect_info_metadata(settings),
+    }
+
+
+@router.post("/api/ai/connect-info/bootstrap", dependencies=[Depends(require_local_operator)])
+@limiter.limit("10/minute")
+async def bootstrap_hmac_secret(request: Request):
+    """Mint and persist the OpenClaw HMAC secret if it isn't already set.
+
+    Idempotent: if a secret already exists, returns ``generated: false``
+    and leaves the existing secret untouched. Never returns the secret
+    value in the response body — the operator calls
+    ``POST /api/ai/connect-info/reveal`` to see it.
+    """
     import secrets
     from services.config import get_settings
 
+    settings = get_settings()
+    existing = str(settings.OPENCLAW_HMAC_SECRET or "").strip()
+    if existing:
+        return {
+            "ok": True,
+            "generated": False,
+            "hmac_secret_set": True,
+            "masked_hmac_secret": _mask_hmac_secret(existing),
+            "detail": "HMAC secret already configured. Use /reveal to see it.",
+        }
+
     new_secret = secrets.token_hex(24)  # 48 chars
     _write_env_value("OPENCLAW_HMAC_SECRET", new_secret)
     get_settings.cache_clear()
 
     return {
         "ok": True,
-        "hmac_secret": new_secret,
-        "detail": "HMAC secret regenerated. Update your OpenClaw agent configuration.",
+        "generated": True,
+        "hmac_secret_set": True,
+        "masked_hmac_secret": _mask_hmac_secret(new_secret),
+        "detail": "HMAC secret generated. Call /reveal to copy it into your OpenClaw config.",
     }
 
 
+@router.post("/api/ai/connect-info/reveal", dependencies=[Depends(require_local_operator)])
+@limiter.limit("10/minute")
+async def reveal_hmac_secret(request: Request):
+    """Return the full HMAC secret in the response body.
+
+    POST (not GET) so the secret never lands in URL history, access logs,
+    or browser visit history. Strict ``Cache-Control: no-store`` headers
+    prevent intermediaries from persisting the response. Returns 404 if
+    no secret has been bootstrapped — the frontend should call
+    ``POST /api/ai/connect-info/bootstrap`` first.
+    """
+    from services.config import get_settings
+
+    settings = get_settings()
+    hmac_secret = str(settings.OPENCLAW_HMAC_SECRET or "").strip()
+    if not hmac_secret:
+        raise HTTPException(
+            404,
+            "No HMAC secret configured. Call POST /api/ai/connect-info/bootstrap first.",
+        )
+    return JSONResponse(
+        content={
+            "ok": True,
+            "hmac_secret": hmac_secret,
+            "masked_hmac_secret": _mask_hmac_secret(hmac_secret),
+        },
+        headers=_NO_STORE_HEADERS,
+    )
+
+
+@router.post("/api/ai/connect-info/regenerate", dependencies=[Depends(require_local_operator)])
+@limiter.limit("5/minute")
+async def regenerate_hmac_secret(request: Request):
+    """Generate a new HMAC secret. Old secret immediately stops working.
+
+    Returns the new secret in the response body — this is the only
+    operation where the full secret travels back through the response,
+    because regenerating IS a deliberate destructive action the operator
+    triggered and they need to see the new value once to update their
+    OpenClaw configuration. Strict ``Cache-Control: no-store`` headers
+    keep it from being persisted by browser caches, proxies, or HAR
+    capture tooling.
+    """
+    import secrets
+    from services.config import get_settings
+
+    new_secret = secrets.token_hex(24)  # 48 chars
+    _write_env_value("OPENCLAW_HMAC_SECRET", new_secret)
+    get_settings.cache_clear()
+
+    return JSONResponse(
+        content={
+            "ok": True,
+            "hmac_secret": new_secret,
+            "masked_hmac_secret": _mask_hmac_secret(new_secret),
+            "detail": "HMAC secret regenerated. Update your OpenClaw agent configuration.",
+        },
+        headers=_NO_STORE_HEADERS,
+    )
+
+
 @router.put("/api/ai/connect-info/access-tier", dependencies=[Depends(require_local_operator)])
 @limiter.limit("10/minute")
 async def set_access_tier(request: Request, body: dict):

backend/routers/health.py

@@ -59,6 +59,12 @@ async def health_check(request: Request):
     # when the SPKI-pinned fallback is in effect. The data plane keeps
     # flowing (this is by design — see ais_proxy.js comments) but observers
     # who care about MITM-protection posture deserve a visible signal.
+    #
+    # Plus connectivity health (added 2026-05-23 when stream.aisstream.io
+    # went fully offline): ``connected`` tells the frontend whether ship
+    # data is actually flowing. When false, a banner explains that ships
+    # are unavailable due to an upstream outage — better than the user
+    # silently seeing an empty ocean and assuming we broke something.
     ais_status: dict = {}
     try:
         from services.ais_stream import ais_proxy_status
@@ -69,6 +75,15 @@ async def health_check(request: Request):
         # Don't override a worse top-level status if SLOs already failed,
         # but escalate ok -> degraded so the field surfaces in dashboards.
         top_status = "degraded"
+    # AIS_API_KEY not configured is "feature off", not "system broken" —
+    # so we only escalate when the operator opted into AIS (key set) AND
+    # the stream is currently offline.
+    if (
+        os.environ.get("AIS_API_KEY")
+        and ais_status.get("connected") is False
+        and top_status == "ok"
+    ):
+        top_status = "degraded"
 
     return {
         "status": top_status,

backend/routers/mesh_peer_sync.py

@@ -85,6 +85,64 @@ async def infonet_peer_push(request: Request):
     return {"ok": True, **result}
 
 
+@router.post("/api/mesh/dm/replicate-envelope")
+@limiter.limit("60/minute")
+async def dm_replicate_envelope(request: Request):
+    """Accept a DM envelope replicated from a peer relay (cross-node mailbox).
+
+    Companion endpoint to ``DMRelay.replicate_to_peers`` (outbound, in
+    ``mesh_dm_relay.py``). The sender's relay POSTs an encrypted DM
+    envelope here after a successful local ``deposit``; this endpoint
+    re-enforces the per-(sender, recipient) anti-spam cap and stores
+    the envelope in the local mailbox if accepted.
+
+    The cap is the network rule: a hostile sender's relay can spool
+    extras locally, but every honest peer enforces the cap on inbound
+    replication. Recipient polling from any honest peer therefore
+    never sees more than ``MESH_DM_PENDING_PER_SENDER_LIMIT`` pending
+    from any one sender, no matter how many spam attempts were tried.
+
+    Same HMAC auth pattern as ``infonet_peer_push`` and ``gate_peer_push``.
+    """
+    content_length = request.headers.get("content-length")
+    if content_length:
+        try:
+            # DM envelopes are bounded by MESH_DM_MAX_MSG_BYTES + envelope
+            # overhead; 64 KB is a generous ceiling.
+            if int(content_length) > 65_536:
+                return Response(
+                    content='{"ok":false,"detail":"Request body too large (max 64KB)"}',
+                    status_code=413, media_type="application/json",
+                )
+        except (ValueError, TypeError):
+            pass
+    body_bytes = await request.body()
+    if not _verify_peer_push_hmac(request, body_bytes):
+        return Response(
+            content='{"ok":false,"detail":"Invalid or missing peer HMAC"}',
+            status_code=403, media_type="application/json",
+        )
+    try:
+        body = json_mod.loads(body_bytes or b"{}")
+    except (ValueError, TypeError):
+        return Response(
+            content='{"ok":false,"detail":"Invalid JSON body"}',
+            status_code=400, media_type="application/json",
+        )
+    envelope = body.get("envelope")
+    if not isinstance(envelope, dict):
+        return {"ok": False, "detail": "envelope must be an object"}
+
+    originating_peer = _peer_hmac_url_from_request(request) or ""
+
+    from services.mesh.mesh_dm_relay import dm_relay
+    result = dm_relay.accept_replica(
+        envelope=envelope,
+        originating_peer_url=originating_peer,
+    )
+    return result
+
+
 @router.post("/api/mesh/gate/peer-push")
 @limiter.limit("30/minute")
 async def gate_peer_push(request: Request):

backend/services/ais_stream.py

@@ -350,19 +350,58 @@ _proxy_process = None
 # path during an upstream cert outage. Surfaced via ais_proxy_status() for
 # /api/health.
 _proxy_status: dict = {}
+# Upstream-connectivity telemetry (added when stream.aisstream.io went fully
+# offline on 2026-05-23). ``_last_msg_at`` is the unix timestamp of the most
+# recent vessel message received from the proxy. ``_proxy_spawn_count`` is
+# how many times we've started the node proxy; combined with no recent
+# messages it tells us the proxy is respawning in a tight loop because the
+# upstream is unreachable. Surfaced via ais_proxy_status() so the operator
+# can see "AIS is dead" instead of guessing whether it's their map filter,
+# their api key, or upstream.
+_last_msg_at: float = 0.0
+_proxy_spawn_count: int = 0
 _VESSEL_TRAIL_INTERVAL_S = 120
 _VESSEL_TRAIL_MAX_POINTS = 240
 
 
-def ais_proxy_status() -> dict:
-    """Return a copy of the latest ais_proxy.js status (issue #258).
+# How stale "last vessel message" can be before we consider the stream
+# disconnected. AISStream typically pushes multiple messages/sec, so a 60s
+# gap means something's wrong upstream or in transit.
+_AIS_CONNECTED_FRESHNESS_S = 60
 
-    Currently surfaces ``degraded_tls`` (bool) which is true when the
-    proxy is using SPKI-pinned fallback because AISStream's cert expired.
-    Returns an empty dict when no status has been received yet.
+
+def ais_proxy_status() -> dict:
+    """Return a copy of the latest ais_proxy.js status + connectivity health.
+
+    Fields:
+      * ``degraded_tls`` (bool, issue #258) — true when the proxy is using
+        SPKI-pinned fallback because AISStream's cert expired.
+      * ``connected`` (bool) — true when we received a vessel message in
+        the last ``_AIS_CONNECTED_FRESHNESS_S`` seconds.
+      * ``last_msg_age_seconds`` (int | None) — seconds since the last
+        vessel message; None if we've never received one.
+      * ``proxy_spawn_count`` (int) — how many times we've spawned the
+        node proxy. Sustained increases here without ``connected`` means
+        we're respawning in a tight loop because upstream is dead.
+
+    Returns an empty dict when called before the AIS subsystem starts
+    (e.g. during tests or when no API key is set).
     """
     with _vessels_lock:
-        return dict(_proxy_status)
+        status = dict(_proxy_status)
+        last = _last_msg_at
+        spawns = _proxy_spawn_count
+
+    now = time.time()
+    if last > 0:
+        last_age = int(now - last)
+        status["last_msg_age_seconds"] = last_age
+        status["connected"] = last_age <= _AIS_CONNECTED_FRESHNESS_S
+    else:
+        status["last_msg_age_seconds"] = None
+        status["connected"] = False
+    status["proxy_spawn_count"] = spawns
+    return status
 
 import os
 
@@ -588,8 +627,10 @@ def _ais_stream_loop():
                 env=proxy_env,
                 **popen_kwargs,
             )
+            global _proxy_spawn_count
             with _vessels_lock:
                 _proxy_process = process
+                _proxy_spawn_count += 1
 
             # Drain stderr in a background thread to prevent deadlock
             import threading
@@ -645,9 +686,15 @@ def _ais_stream_loop():
                 if not mmsi:
                     continue
 
+                # Telemetry: stamp the timestamp of the most recent real
+                # vessel message. ais_proxy_status() reads this to decide
+                # whether the stream is currently "connected" — i.e. has
+                # any data flowed in the last 60s.
+                global _last_msg_at
                 with _vessels_lock:
+                    _last_msg_at = time.time()
                     if mmsi not in _vessels:
-                        _vessels[mmsi] = {"_updated": time.time()}
+                        _vessels[mmsi] = {"_updated": _last_msg_at}
                     vessel = _vessels[mmsi]
 
                 # Update position from PositionReport or StandardClassBPositionReport

backend/services/config.py

@@ -116,6 +116,21 @@ class Settings(BaseSettings):
     MESH_DM_REQUEST_MAILBOX_LIMIT: int = 12
     MESH_DM_SHARED_MAILBOX_LIMIT: int = 48
     MESH_DM_SELF_MAILBOX_LIMIT: int = 12
+    # Anti-spam: cap on distinct UNACKED messages a single sender can have
+    # parked in a single recipient's mailbox at any one time. Once the
+    # recipient pulls (acks) a message, the sender's quota for that pair
+    # frees up. Default 2 — a sender who wants to deliver more must wait
+    # for the recipient to actually read the prior messages.
+    #
+    # This cap is enforced TWICE: once on the local deposit path (the
+    # sender's own node refuses to spool the 3rd message) AND once on
+    # the replication-acceptance path (honest peer relays refuse to
+    # accept inbound replicas that would put them over the cap). The
+    # double enforcement makes the rule a NETWORK rule — patching out
+    # the local check on a hostile sender's relay doesn't let extras
+    # propagate, because every honest peer enforces the same cap on
+    # inbound replication.
+    MESH_DM_PENDING_PER_SENDER_LIMIT: int = 2
     MESH_BLOCK_LEGACY_AGENT_ID_LOOKUP: bool = True
     MESH_ALLOW_COMPAT_DM_INVITE_IMPORT: bool = False
     MESH_ALLOW_COMPAT_DM_INVITE_IMPORT_UNTIL: str = ""

backend/services/constants.py

@@ -11,8 +11,13 @@ DEFAULT_TRAIL_TTL_S = 300  # 5 min - trail TTL for non-tracked flights
 HOLD_PATTERN_DEGREES = 300  # Total heading change to flag holding pattern
 GPS_JAMMING_NACP_THRESHOLD = 8  # NACp below this = degraded GPS signal
 GPS_JAMMING_GRID_SIZE = 1.0  # 1 degree grid for aggregation
-GPS_JAMMING_MIN_RATIO = 0.30  # 30% degraded aircraft to flag zone
-GPS_JAMMING_MIN_AIRCRAFT = 5  # Min aircraft in grid cell for statistical significance
+# Tuned 2026-05: previously 0.30 / 5 aircraft which — combined with the
+# -1 noise cushion in the detector AND the pre-fix nac_p==0 filter that
+# discarded jamming victims — meant the layer almost never lit up.
+# Lowering the bar so genuine jamming zones with sparser ADS-B coverage
+# clear (eastern Med, Russia/Ukraine border, Iran/Iraq).
+GPS_JAMMING_MIN_RATIO = 0.20  # 20% degraded aircraft to flag zone
+GPS_JAMMING_MIN_AIRCRAFT = 3  # Min aircraft in grid cell for statistical significance
 
 # ─── Network & Circuit Breaker ──────────────────────────────────────────────
 CIRCUIT_BREAKER_TTL_S = 120  # Skip domain for 2 min after total failure

backend/services/data_fetcher.py

@@ -777,6 +777,39 @@ def start_scheduler():
         misfire_grace_time=60,
     )
 
+    # Flight observation pruning — drops icao24 → first_seen_at entries we
+    # haven't seen in an hour. Same cadence as AIS prune for symmetry; the
+    # per-tick scan is O(in-flight aircraft) so it's cheap.
+    from services.fetchers.flight_observations import prune as _prune_flight_observations
+    _scheduler.add_job(
+        lambda: _run_task_with_health(_prune_flight_observations, "prune_flight_observations"),
+        "interval",
+        minutes=5,
+        id="flight_observation_prune",
+        max_instances=1,
+        misfire_grace_time=60,
+    )
+
+    # AISHub REST fallback — slow polling when the AISStream WebSocket
+    # primary is offline. Configurable interval via
+    # AISHUB_POLL_INTERVAL_MINUTES env (default 20 min). Operator must
+    # set AISHUB_USERNAME to opt in. The fetcher is gated internally on
+    # the primary being disconnected, so this job is cheap when the
+    # WebSocket is healthy (early-returns after a status check).
+    from services.fetchers.aishub_fallback import (
+        aishub_poll_interval_minutes,
+        fetch_aishub_vessels,
+    )
+    _aishub_interval = aishub_poll_interval_minutes()
+    _scheduler.add_job(
+        lambda: _run_task_with_health(fetch_aishub_vessels, "fetch_aishub_vessels"),
+        "interval",
+        minutes=_aishub_interval,
+        id="aishub_fallback",
+        max_instances=1,
+        misfire_grace_time=120,
+    )
+
     # Route database — bulk refresh from vrs-standing-data.adsb.lol every 5
     # days. Replaces the legacy /api/0/routeset POST (blocked under our UA,
     # and broken upstream). Airline schedules change on a quarterly cycle,
@@ -960,16 +993,19 @@ def start_scheduler():
         misfire_grace_time=600,
     )
 
-    # UAP sightings (NUFORC) — daily at 12:00 UTC
+    # UAP sightings (NUFORC) — weekly on Mondays at 12:00 UTC. The layer is a
+    # rolling last-60-days digest; refreshing once a week is enough cadence
+    # for human-readable map exploration and keeps load on nuforc.org light.
     _scheduler.add_job(
         lambda: _run_task_with_health(
             lambda: fetch_uap_sightings(force_refresh=True),
             "fetch_uap_sightings",
         ),
         "cron",
+        day_of_week="mon",
         hour=12,
         minute=0,
-        id="uap_sightings_daily",
+        id="uap_sightings_weekly",
         max_instances=1,
         misfire_grace_time=3600,
     )

backend/services/fetchers/aishub_fallback.py

@@ -0,0 +1,290 @@
+"""AISHub REST fallback for ship tracking when AISStream is unreachable.
+
+Background
+----------
+On 2026-05-23 ``stream.aisstream.io`` (the primary live AIS WebSocket feed)
+went fully offline. Backend's only ship signal vanished. This module polls
+``data.aishub.net``'s free REST API on a slow cadence (default 20 min) when
+the WebSocket primary is disconnected, so the ships layer doesn't go fully
+dark during upstream outages.
+
+Why 20 minutes
+--------------
+AISHub's free tier is rate-limited and explicitly asks consumers to be
+courteous. 20 minutes is well inside their limits, gives ships time to
+move enough to look "alive" on the map, and won't drain their service.
+Configurable via the ``AISHUB_POLL_INTERVAL_MINUTES`` env var (clamped to
+[1, 360]).
+
+Why slow vs primary
+-------------------
+This is degraded mode, not a replacement. A ship at 20 knots moves about
+6 nautical miles in 20 minutes — visible on the map but coarser than the
+real-time WebSocket signal. When AISStream comes back online, the
+WebSocket data will overwrite these records via the same ``_vessels``
+dict and ``source`` will flip from ``"aishub"`` back to upstream-live.
+
+Opt-in
+------
+Operator must set ``AISHUB_USERNAME`` (free registration at
+https://www.aishub.net/api). If unset, this fetcher is a no-op.
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+import os
+import time
+from typing import Any
+
+from services.network_utils import fetch_with_curl
+
+logger = logging.getLogger(__name__)
+
+
+AISHUB_URL = "https://data.aishub.net/ws.php"
+
+
+def aishub_username() -> str:
+    return str(os.environ.get("AISHUB_USERNAME", "")).strip()
+
+
+def aishub_fallback_enabled() -> bool:
+    """Returns True only when the operator has registered with AISHub and
+    set ``AISHUB_USERNAME``. The presence of the username is the opt-in."""
+    return bool(aishub_username())
+
+
+def aishub_poll_interval_minutes() -> int:
+    """Default 20 minutes. Clamped to [1, 360] so a hostile or
+    misconfigured env var can't either hammer the upstream or silence the
+    fallback for a day."""
+    raw = os.environ.get("AISHUB_POLL_INTERVAL_MINUTES", "20")
+    try:
+        value = int(str(raw).strip())
+    except (TypeError, ValueError):
+        value = 20
+    return max(1, min(360, value))
+
+
+def _should_run_fallback() -> bool:
+    """Only run when the primary WebSocket is disconnected. Avoids stomping
+    over fresher live data when AISStream is healthy.
+
+    Returns False if:
+      * AISHub isn't configured (no username)
+      * AISStream primary is currently connected (recent vessel messages)
+
+    Returns True only when AIS is configured-but-down. The
+    ``proxy_spawn_count > 0`` guard means "the primary has at least tried
+    to run" — if the user set AISHUB_USERNAME but not AIS_API_KEY at all,
+    AISHub will still serve as a primary on its own slow cadence.
+    """
+    if not aishub_fallback_enabled():
+        return False
+    try:
+        from services.ais_stream import ais_proxy_status
+        status = ais_proxy_status() or {}
+    except Exception:
+        return True  # ais_stream not importable? still try AISHub.
+    # If the WebSocket primary is connected, skip the fallback — fresher
+    # data is already flowing.
+    if status.get("connected") is True:
+        return False
+    return True
+
+
+def _parse_aishub_response(payload: str) -> list[dict]:
+    """Parse the AISHub JSON response into a list of vessel records.
+
+    Successful response shape::
+
+        [
+            {"ERROR": false, "USERNAME": "...", "FORMAT": "1", "RECORDS": N},
+            [{"MMSI": ..., "LATITUDE": ..., "LONGITUDE": ..., ...}, ...]
+        ]
+
+    Error response shape::
+
+        [{"ERROR": true, "ERROR_MESSAGE": "..."}]
+
+    Empty payload (e.g. silent rate-limit drop) returns ``[]``.
+    """
+    if not payload or not payload.strip():
+        return []
+    try:
+        data = json.loads(payload)
+    except json.JSONDecodeError as e:
+        logger.warning("AISHub: response is not JSON: %s", e)
+        return []
+    if not isinstance(data, list) or not data:
+        return []
+    header = data[0] if isinstance(data[0], dict) else {}
+    if header.get("ERROR") is True:
+        logger.warning(
+            "AISHub: upstream error: %s",
+            header.get("ERROR_MESSAGE", "<unspecified>"),
+        )
+        return []
+    if len(data) < 2 or not isinstance(data[1], list):
+        return []
+    return [row for row in data[1] if isinstance(row, dict)]
+
+
+def _normalize_record(row: dict) -> dict | None:
+    """Map an AISHub vessel record to our internal vessel schema.
+
+    Returns None when the record can't be used (no MMSI, bad position,
+    sentinel "not available" lat/lng).
+    """
+    try:
+        mmsi = int(row.get("MMSI") or 0)
+    except (TypeError, ValueError):
+        return None
+    if not mmsi:
+        return None
+    try:
+        lat = float(row.get("LATITUDE"))
+        lng = float(row.get("LONGITUDE"))
+    except (TypeError, ValueError):
+        return None
+    # AIS uses 91/181 as "no position available" sentinels.
+    if abs(lat) > 90 or abs(lng) > 180:
+        return None
+    if lat == 91.0 or lng == 181.0:
+        return None
+    # SOG raw 102.3 is "speed not available"; sanitize to 0.
+    try:
+        sog_raw = float(row.get("SOG") or 0)
+    except (TypeError, ValueError):
+        sog_raw = 0.0
+    sog = 0.0 if sog_raw >= 102.2 else sog_raw
+    try:
+        cog = float(row.get("COG") or 0)
+    except (TypeError, ValueError):
+        cog = 0.0
+    try:
+        heading_raw = int(row.get("HEADING") or 511)
+    except (TypeError, ValueError):
+        heading_raw = 511
+    # AIS heading sentinel 511 = "not available" — fall back to COG.
+    heading = heading_raw if heading_raw != 511 else cog
+    try:
+        ais_type = int(row.get("TYPE") or 0)
+    except (TypeError, ValueError):
+        ais_type = 0
+    return {
+        "mmsi": mmsi,
+        "lat": lat,
+        "lng": lng,
+        "sog": sog,
+        "cog": cog,
+        "heading": heading,
+        "name": str(row.get("NAME") or "").strip() or "UNKNOWN",
+        "callsign": str(row.get("CALLSIGN") or "").strip(),
+        "destination": str(row.get("DEST") or "").strip().replace("@", "") or "",
+        "imo": int(row.get("IMO") or 0),
+        "ais_type_code": ais_type,
+    }
+
+
+def fetch_aishub_vessels() -> int:
+    """Poll AISHub and merge vessels into the shared ``_vessels`` store.
+
+    Returns the number of vessels updated (0 on skip, error, or no data).
+    Designed to be called by the APScheduler tier — see
+    ``data_fetcher.py`` for the 20-minute interval job that wraps this.
+    """
+    if not _should_run_fallback():
+        logger.debug("AISHub fallback skipped: primary connected or not configured")
+        return 0
+
+    username = aishub_username()
+    url = (
+        f"{AISHUB_URL}?username={username}&format=1&output=json"
+        f"&compress=0"
+    )
+
+    try:
+        response = fetch_with_curl(url, timeout=30)
+    except Exception as e:
+        logger.warning("AISHub fetch failed: %s", e)
+        return 0
+
+    if not response or response.status_code != 200:
+        logger.warning(
+            "AISHub HTTP %s",
+            getattr(response, "status_code", "None"),
+        )
+        return 0
+
+    rows = _parse_aishub_response(getattr(response, "text", "") or "")
+    if not rows:
+        return 0
+
+    # Inline imports to avoid a circular dependency at module load time
+    # (ais_stream imports lots of things and is loaded by main.py).
+    from services.ais_stream import (
+        _vessels,
+        _vessels_lock,
+        _record_vessel_trail_locked,
+        classify_vessel,
+        get_country_from_mmsi,
+    )
+
+    now = time.time()
+    count = 0
+    with _vessels_lock:
+        for row in rows:
+            normalized = _normalize_record(row)
+            if normalized is None:
+                continue
+            mmsi = normalized["mmsi"]
+            vessel = _vessels.setdefault(mmsi, {"mmsi": mmsi})
+            # Don't overwrite fresher live data: if the WebSocket pushed an
+            # update for this MMSI more recently than now-1s (race during
+            # the brief reconnection window) keep the live one.
+            last = float(vessel.get("_updated") or 0)
+            if last > now - 1:
+                continue
+            vessel.update(
+                {
+                    "lat": normalized["lat"],
+                    "lng": normalized["lng"],
+                    "sog": normalized["sog"],
+                    "cog": normalized["cog"],
+                    "heading": normalized["heading"],
+                    "_updated": now,
+                    "source": "aishub",
+                }
+            )
+            if normalized["name"] and normalized["name"] != "UNKNOWN":
+                vessel["name"] = normalized["name"]
+            if normalized["callsign"]:
+                vessel["callsign"] = normalized["callsign"]
+            if normalized["destination"]:
+                vessel["destination"] = normalized["destination"]
+            if normalized["imo"]:
+                vessel["imo"] = normalized["imo"]
+            if normalized["ais_type_code"]:
+                vessel["ais_type_code"] = normalized["ais_type_code"]
+                vessel["type"] = classify_vessel(normalized["ais_type_code"], mmsi)
+            if not vessel.get("country"):
+                vessel["country"] = get_country_from_mmsi(mmsi)
+            _record_vessel_trail_locked(
+                mmsi,
+                normalized["lat"],
+                normalized["lng"],
+                normalized["sog"],
+                now,
+            )
+            count += 1
+
+    if count:
+        logger.info(
+            "AISHub fallback: merged %d vessels (poll interval %d min)",
+            count,
+            aishub_poll_interval_minutes(),
+        )
+    return count

backend/services/fetchers/earth_observation.py

@@ -1383,10 +1383,21 @@ def _build_uap_sightings_from_hf_mirror() -> list[dict]:
     This is a resilience fallback for local/Windows runs where nuforc.org is
     Cloudflare-gated and the Mapbox token is not configured. It is not as fresh
     as the live NUFORC AJAX feed, but it keeps the layer visible and cached.
+
+    Date-cutoff guard: the kcimc/NUFORC HF dataset is a static snapshot whose
+    maintainer refreshes it sporadically. Without a cutoff, sorting by
+    occurred-desc and taking the top N rows returns whatever the mirror's
+    newest rows happen to be — which can be years old if the snapshot is
+    stale. We apply the same ``_NUFORC_RECENT_DAYS`` window the live path
+    uses (60 days). If the HF mirror has nothing inside the window we return
+    ``[]`` rather than silently serving 3-year-old "newest" rows.
     """
     from services.fetchers.nuforc_enrichment import _HF_CSV_URL, _parse_date
     from services.geocode_validate import coord_in_country
 
+    cutoff_dt = datetime.utcnow() - timedelta(days=_NUFORC_RECENT_DAYS)
+    cutoff_str = cutoff_dt.strftime("%Y-%m-%d")
+
     try:
         response = fetch_with_curl(_HF_CSV_URL, timeout=180, follow_redirects=True)
         if not response or response.status_code != 200:
@@ -1400,6 +1411,7 @@ def _build_uap_sightings_from_hf_mirror() -> list[dict]:
         return []
 
     candidates: list[dict] = []
+    stale_rows_dropped = 0
     try:
         reader = csv.DictReader(io.StringIO(response.text))
         for row in reader:
@@ -1410,6 +1422,9 @@ def _build_uap_sightings_from_hf_mirror() -> list[dict]:
             )
             if not occurred:
                 continue
+            if occurred < cutoff_str:
+                stale_rows_dropped += 1
+                continue
             raw_location = _normalize_uap_location(
                 row.get("Location", "")
                 or row.get("City", "")
@@ -1444,6 +1459,19 @@ def _build_uap_sightings_from_hf_mirror() -> list[dict]:
         logger.warning("UAP sightings: HF fallback parse failed: %s", e)
         return []
 
+    if not candidates:
+        # HF mirror returned rows, but none inside the rolling window. This is
+        # the smoking gun for "the public HF dataset hasn't been refreshed in
+        # years" — log loudly so the operator sees it instead of guessing.
+        logger.error(
+            "UAP sightings: HF fallback yielded 0 rows within last %d days "
+            "(dropped %d stale rows). HF mirror is likely stale; the layer "
+            "will be empty until the live NUFORC path recovers.",
+            _NUFORC_RECENT_DAYS,
+            stale_rows_dropped,
+        )
+        return []
+
     candidates.sort(key=lambda row: (row["occurred"], row["posted"], row["id"]), reverse=True)
     candidates = candidates[:_NUFORC_HF_FALLBACK_LIMIT]
 
@@ -1515,13 +1543,29 @@ def fetch_uap_sightings(*, force_refresh: bool = False):
 
     sightings = _load_nuforc_sightings_cache(force_refresh=force_refresh)
     if sightings is None:
+        live_error: Exception | None = None
         try:
             sightings = _build_recent_uap_sightings()
         except Exception as e:
+            live_error = e
             logger.warning("UAP sightings: live NUFORC rebuild failed, using fallback: %s", e)
             sightings = _build_uap_sightings_from_hf_mirror()
         if sightings:
             _save_nuforc_sightings_cache(sightings)
+        elif live_error is not None:
+            # Both paths failed: live raised AND HF fallback returned empty
+            # (either the HF mirror is stale beyond the cutoff or the network
+            # is gone entirely). The previous code silently set the layer to
+            # ``[]`` and kept marking it fresh; that masked the failure for
+            # days. Surface it via assert_canary so the health registry shows
+            # the layer as broken instead of "fresh and empty".
+            from services.slo import assert_canary
+            assert_canary("uap_sightings", 0)
+            logger.error(
+                "UAP sightings: both live NUFORC and HF fallback produced 0 "
+                "rows; layer is unavailable. Live error: %s",
+                live_error,
+            )
 
     with _data_lock:
         latest_data["uap_sightings"] = sightings or []

backend/services/fetchers/flight_observations.py

@@ -0,0 +1,148 @@
+"""Per-aircraft observation tracking for cumulative fuel/CO2 estimates.
+
+Background
+----------
+The pre-existing emissions enrichment attached a *rate* to each flight
+(GPH and kg/hr) based on aircraft model. Users — reasonably — wanted the
+running total: how much fuel HAS this plane burned since we started
+seeing it? Multiplying the rate by elapsed observation time gets us
+there, but it requires somewhere to remember "when did this icao24
+first appear on our radar?"
+
+Why this lives outside ``flight_trails``
+----------------------------------------
+``flight_trails`` is sized and pruned aggressively for map rendering
+(5-minute TTL for untracked aircraft, 200 trail points max). That's
+wrong for cumulative burn: if a plane has been airborne 2 hours but
+its trail was pruned 30 min in, the "first trail point" timestamp is
+30 min ago, not 2h ago. Worse, when the trail expires and re-creates,
+the cumulative counter would reset mid-flight.
+
+This module tracks observation lifecycle separately:
+
+* When a hex is first observed: start a new flight session.
+* While observed regularly (gap < ``REOPEN_GAP_S``): keep accumulating.
+* When unseen for longer than ``REOPEN_GAP_S``: treat next sighting as
+  a new session (the plane landed and took off again, or it's a
+  different leg). Reset ``first_seen_at``.
+* Stale sessions are pruned every ``PRUNE_INTERVAL_S`` so memory stays
+  bounded.
+
+The user explicitly asked for this counting semantic: "as soon as a
+plane appears there should be a counter that keeps a running count of
+the fuel being burned... If there is no estimate take off time then it
+can just be from the time the server starts to keep a log of whats in
+the air."
+"""
+
+from __future__ import annotations
+
+import threading
+import time
+
+
+# Gap between sightings that resets the session. ADS-B refreshes the
+# whole aircraft list every minute or two, so anything over a few
+# minutes means the plane left our coverage window (landed, transit
+# through dead zone, etc). 15 minutes is conservative.
+REOPEN_GAP_S = 15 * 60
+
+# Don't accumulate runaway memory: drop entries unseen for an hour.
+PRUNE_AFTER_S = 60 * 60
+
+# Cap on accumulated airtime per session so a single bug elsewhere
+# (e.g. ts clock skew) can't produce comically large numbers.
+MAX_SESSION_SECONDS = 24 * 3600  # 24h — longest realistic civilian leg
+
+
+_observations: dict[str, dict[str, float]] = {}
+_lock = threading.Lock()
+_last_prune_at = 0.0
+
+
+def record_observation(icao_hex: str, *, now: float | None = None) -> int:
+    """Record a sighting of ``icao_hex`` and return airtime so far (seconds).
+
+    Returns 0 for the first-ever sighting (no elapsed time yet) or when
+    ``icao_hex`` is falsy. The caller can multiply the returned seconds
+    by ``rate_per_hour / 3600`` to get cumulative consumption.
+    """
+    if not icao_hex:
+        return 0
+    key = str(icao_hex).strip().lower()
+    if not key:
+        return 0
+    current = float(now if now is not None else time.time())
+
+    with _lock:
+        entry = _observations.get(key)
+        if entry is None:
+            _observations[key] = {"first_seen_at": current, "last_seen_at": current}
+            return 0
+        # Use explicit ``is None`` checks instead of ``or`` short-circuit:
+        # ``0.0`` is a legitimate timestamp value (e.g. test fixtures
+        # seeding a far-past first_seen_at to exercise the clamp) but
+        # ``0.0 or fallback`` collapses to ``fallback`` because 0.0 is
+        # falsy. Bit me on my own test — leaving the safer form here.
+        last_raw = entry.get("last_seen_at")
+        last_seen = float(last_raw) if last_raw is not None else current
+        gap = current - last_seen
+        if gap > REOPEN_GAP_S:
+            # Treat as a new flight session — the plane landed/disappeared
+            # long enough that the prior cumulative count is no longer
+            # the same flight.
+            _observations[key] = {"first_seen_at": current, "last_seen_at": current}
+            return 0
+        first_raw = entry.get("first_seen_at")
+        first = float(first_raw) if first_raw is not None else current
+        # Clamp absurd values from clock skew or bad input.
+        elapsed = max(0, min(int(current - first), MAX_SESSION_SECONDS))
+        entry["last_seen_at"] = current
+        return elapsed
+
+
+def prune(*, now: float | None = None) -> int:
+    """Drop entries we haven't seen in ``PRUNE_AFTER_S`` seconds.
+
+    Returns number of entries dropped. Safe to call from a scheduler tick;
+    cheap (single dict scan) so cadence doesn't matter much.
+    """
+    current = float(now if now is not None else time.time())
+    dropped = 0
+    with _lock:
+        stale_keys = []
+        for k, v in _observations.items():
+            last_raw = v.get("last_seen_at")
+            last = float(last_raw) if last_raw is not None else 0.0
+            if current - last > PRUNE_AFTER_S:
+                stale_keys.append(k)
+        for k in stale_keys:
+            del _observations[k]
+            dropped += 1
+    return dropped
+
+
+def get_session_seconds(icao_hex: str, *, now: float | None = None) -> int:
+    """Read-only accessor: airtime for a known icao without bumping last-seen.
+
+    Used by tests and external consumers (e.g. when rendering a snapshot
+    of all in-flight aircraft, you want the current value, not to update
+    last_seen_at as a side effect).
+    """
+    if not icao_hex:
+        return 0
+    key = str(icao_hex).strip().lower()
+    with _lock:
+        entry = _observations.get(key)
+        if entry is None:
+            return 0
+        current = float(now if now is not None else time.time())
+        first_raw = entry.get("first_seen_at")
+        first = float(first_raw) if first_raw is not None else current
+        return max(0, min(int(current - first), MAX_SESSION_SECONDS))
+
+
+def _reset_for_tests() -> None:
+    """Drop all observations. Test helper only."""
+    with _lock:
+        _observations.clear()

backend/services/fetchers/flights.py

@@ -17,6 +17,7 @@ from services.network_utils import fetch_with_curl
 from services.fetchers._store import latest_data, _data_lock, _mark_fresh
 from services.fetchers.plane_alert import enrich_with_plane_alert, enrich_with_tracked_names
 from services.fetchers.emissions import get_emissions_info
+from services.fetchers.flight_observations import record_observation as _record_flight_observation
 from services.fetchers.retry import with_retry
 from services.fetchers.route_database import lookup_route
 from services.fetchers.aircraft_database import lookup_aircraft_type
@@ -29,6 +30,88 @@ _RE_AIRLINE_CODE_1 = re.compile(r"^([A-Z]{3})\d")
 _RE_AIRLINE_CODE_2 = re.compile(r"^([A-Z]{3})[A-Z\d]")
 
 
+def detect_gps_jamming_zones(
+    raw_flights: list[dict],
+    *,
+    min_aircraft: int | None = None,
+    min_ratio: float | None = None,
+    nacp_threshold: int | None = None,
+) -> list[dict]:
+    """Detect GPS interference zones from a snapshot of raw ADS-B aircraft.
+
+    Methodology mirrors GPSJam.org / Flightradar24: bin aircraft into 1°x1°
+    grid cells, flag cells where the fraction of aircraft reporting degraded
+    NACp clears a threshold.
+
+    Inputs
+    ------
+    raw_flights:
+        Iterable of dicts. Each item is expected to carry ``lat``, ``lng``
+        (or ``lon``), and ``nac_p``. Records missing position OR missing
+        ``nac_p`` entirely (typical for OpenSky-sourced flights) are
+        skipped — absence-of-data isn't evidence of anything.
+
+    nac_p == 0 IS counted as degraded. Pre-fix code skipped it on the theory
+    that "0 = old transponder, never computed accuracy." That's only half
+    right: modern Mode-S Enhanced Surveillance transponders also fall back
+    to nac_p=0 when they lose GPS lock entirely — which is exactly the
+    jamming signature we're trying to detect. Filtering 0 out was discarding
+    the strongest evidence.
+
+    Denoising:
+        1. Require ``min_aircraft`` per grid cell for statistical validity.
+        2. Subtract 1 from degraded count per cell (GPSJam's technique) so
+           a single quirky transponder can't flag an entire zone.
+        3. Require ratio ``adjusted_degraded / total > min_ratio``.
+
+    All thresholds default to the module-level constants but can be
+    overridden for testing.
+    """
+    min_aircraft = GPS_JAMMING_MIN_AIRCRAFT if min_aircraft is None else int(min_aircraft)
+    min_ratio = GPS_JAMMING_MIN_RATIO if min_ratio is None else float(min_ratio)
+    nacp_threshold = (
+        GPS_JAMMING_NACP_THRESHOLD if nacp_threshold is None else int(nacp_threshold)
+    )
+
+    jamming_grid: dict[str, dict[str, int]] = {}
+    for rf in raw_flights or []:
+        rlat = rf.get("lat")
+        rlng = rf.get("lng") if rf.get("lng") is not None else rf.get("lon")
+        if rlat is None or rlng is None:
+            continue
+        nacp = rf.get("nac_p")
+        if nacp is None:
+            continue
+        grid_key = f"{int(rlat)},{int(rlng)}"
+        cell = jamming_grid.setdefault(grid_key, {"degraded": 0, "total": 0})
+        cell["total"] += 1
+        if nacp < nacp_threshold:
+            cell["degraded"] += 1
+
+    jamming_zones: list[dict] = []
+    for gk, counts in jamming_grid.items():
+        if counts["total"] < min_aircraft:
+            continue
+        adjusted_degraded = max(counts["degraded"] - 1, 0)
+        if adjusted_degraded == 0:
+            continue
+        ratio = adjusted_degraded / counts["total"]
+        if ratio > min_ratio:
+            lat_i, lng_i = gk.split(",")
+            severity = "low" if ratio < 0.5 else "medium" if ratio < 0.75 else "high"
+            jamming_zones.append(
+                {
+                    "lat": int(lat_i) + 0.5,
+                    "lng": int(lng_i) + 0.5,
+                    "severity": severity,
+                    "ratio": round(ratio, 2),
+                    "degraded": counts["degraded"],
+                    "total": counts["total"],
+                }
+            )
+    return jamming_zones
+
+
 # ---------------------------------------------------------------------------
 # OpenSky Network API Client (OAuth2)
 # ---------------------------------------------------------------------------
@@ -459,6 +542,18 @@ def _classify_and_publish(all_adsb_flights):
 
             ac_category = "heli" if model_upper in _HELI_TYPES_BACKEND else "plane"
 
+            # Source attribution: prefer the explicit ``source`` tag stamped
+            # at fetch time (adsb.lol, OpenSky). If absent, fall back to the
+            # legacy ``supplemental_source`` (airplanes.live, adsb.fi) so
+            # supplementals are still attributed without changing their
+            # tagger. Final fallback "adsb.lol" preserves prior behavior for
+            # any caller that synthesizes records without going through one
+            # of our fetchers (e.g. tests).
+            source = (
+                f.get("source")
+                or f.get("supplemental_source")
+                or "adsb.lol"
+            )
             flights.append(
                 {
                     "callsign": flight_str,
@@ -480,6 +575,7 @@ def _classify_and_publish(all_adsb_flights):
                     "airline_code": airline_code,
                     "aircraft_category": ac_category,
                     "nac_p": f.get("nac_p"),
+                    "source": source,
                 }
             )
         except (ValueError, TypeError, KeyError, AttributeError) as loop_e:
@@ -506,6 +602,22 @@ def _classify_and_publish(all_adsb_flights):
         if model:
             emi = get_emissions_info(model)
             if emi:
+                # Cumulative fuel/CO2: multiply the per-hour rate by how
+                # long we've been observing this airframe. Users want to
+                # see the *amount* burned, not just the rate. If we've
+                # never seen this hex before, observed_seconds is 0 and
+                # the cumulative values are 0 until the next refresh —
+                # the rate is still useful info on its own.
+                observed_seconds = _record_flight_observation(
+                    f.get("icao24") or ""
+                )
+                elapsed_h = observed_seconds / 3600.0
+                emi = {
+                    **emi,
+                    "observed_seconds": observed_seconds,
+                    "fuel_gallons_burned": round(emi["fuel_gph"] * elapsed_h, 1),
+                    "co2_kg_emitted": round(emi["co2_kg_per_hour"] * elapsed_h, 1),
+                }
                 f["emissions"] = emi
 
         callsign = f.get("callsign", "").strip().upper()
@@ -724,56 +836,8 @@ def _classify_and_publish(all_adsb_flights):
         latest_data["military_flights"] = military_snapshot
 
     # --- GPS Jamming Detection ---
-    # Uses NACp (Navigation Accuracy Category – Position) from ADS-B to infer
-    # GPS interference zones, similar to GPSJam.org / Flightradar24.
-    # NACp < 8 = position accuracy worse than the FAA-mandated 0.05 NM.
-    #
-    # Denoising (to suppress false positives from old GA transponders):
-    # 1. Skip nac_p == 0 ("unknown accuracy") — old transponders that never
-    #    computed accuracy, NOT evidence of jamming.  Real jamming shows 1-7.
-    # 2. Require minimum aircraft per grid cell for statistical validity.
-    # 3. Subtract 1 from degraded count per cell (GPSJam's technique) so a
-    #    single quirky transponder can't flag an entire zone.
-    # 4. Require the adjusted ratio to exceed the threshold.
     try:
-        jamming_grid = {}
-        raw_flights = raw_flights_snapshot
-        for rf in raw_flights:
-            rlat = rf.get("lat")
-            rlng = rf.get("lng") or rf.get("lon")
-            if rlat is None or rlng is None:
-                continue
-            nacp = rf.get("nac_p")
-            if nacp is None or nacp == 0:
-                continue
-            grid_key = f"{int(rlat)},{int(rlng)}"
-            if grid_key not in jamming_grid:
-                jamming_grid[grid_key] = {"degraded": 0, "total": 0}
-            jamming_grid[grid_key]["total"] += 1
-            if nacp < GPS_JAMMING_NACP_THRESHOLD:
-                jamming_grid[grid_key]["degraded"] += 1
-
-        jamming_zones = []
-        for gk, counts in jamming_grid.items():
-            if counts["total"] < GPS_JAMMING_MIN_AIRCRAFT:
-                continue
-            adjusted_degraded = max(counts["degraded"] - 1, 0)
-            if adjusted_degraded == 0:
-                continue
-            ratio = adjusted_degraded / counts["total"]
-            if ratio > GPS_JAMMING_MIN_RATIO:
-                lat_i, lng_i = gk.split(",")
-                severity = "low" if ratio < 0.5 else "medium" if ratio < 0.75 else "high"
-                jamming_zones.append(
-                    {
-                        "lat": int(lat_i) + 0.5,
-                        "lng": int(lng_i) + 0.5,
-                        "severity": severity,
-                        "ratio": round(ratio, 2),
-                        "degraded": counts["degraded"],
-                        "total": counts["total"],
-                    }
-                )
+        jamming_zones = detect_gps_jamming_zones(raw_flights_snapshot)
         with _data_lock:
             latest_data["gps_jamming"] = jamming_zones
         if jamming_zones:
@@ -849,7 +913,15 @@ def _fetch_adsb_lol_regions():
             res = fetch_with_curl(url, timeout=10)
             if res.status_code == 200:
                 data = res.json()
-                return data.get("ac", [])
+                aircraft = data.get("ac", [])
+                # Stamp the source at the fetch site so attribution survives
+                # the OpenSky/supplemental dedupe-by-hex merge downstream.
+                # Previously adsb.lol records carried no marker while OpenSky
+                # records got ``is_opensky: True`` — which made flight tooltips
+                # look like everything came from OpenSky.
+                for a in aircraft:
+                    a["source"] = "adsb.lol"
+                return aircraft
         except (
             requests.RequestException,
             ConnectionError,
@@ -932,6 +1004,7 @@ def _enrich_with_opensky_and_supplemental(adsb_flights):
                                     "gs": (s[9] * 1.94384) if s[9] else 0,
                                     "t": "Unknown",
                                     "is_opensky": True,
+                                    "source": "OpenSky",
                                 }
                             )
                     elif os_res.status_code == 429:

backend/services/fetchers/military.py

@@ -7,6 +7,7 @@ import requests
 from services.network_utils import fetch_with_curl
 from services.fetchers._store import latest_data, _data_lock, _mark_fresh
 from services.fetchers.emissions import get_emissions_info
+from services.fetchers.flight_observations import record_observation as _record_flight_observation
 from services.fetchers.plane_alert import enrich_with_plane_alert
 
 logger = logging.getLogger("services.data_fetcher")
@@ -171,6 +172,7 @@ def fetch_military_flights():
                 h = a.get("hex", "").lower()
                 if h and h not in seen_hex:
                     seen_hex.add(h)
+                    a["source"] = "adsb.lol"
                     all_mil_ac.append(a)
     except Exception as e:
         logger.warning(f"adsb.lol mil fetch failed: {e}")
@@ -182,6 +184,7 @@ def fetch_military_flights():
                 h = a.get("hex", "").lower()
                 if h and h not in seen_hex:
                     seen_hex.add(h)
+                    a["source"] = "airplanes.live"
                     all_mil_ac.append(a)
             logger.info(f"airplanes.live mil: +{len(resp2.json().get('ac', []))} raw, {len(all_mil_ac)} total unique")
     except Exception as e:
@@ -234,6 +237,7 @@ def fetch_military_flights():
                             "registration": f.get("r", "N/A"),
                             "icao24": icao_hex,
                             "squawk": f.get("squawk", ""),
+                            "source": f.get("source") or "adsb.lol",
                         })
                         continue
 
@@ -258,7 +262,8 @@ def fetch_military_flights():
                         "model": f.get("t", "Unknown"),
                         "icao24": icao_hex,
                         "speed_knots": speed_knots,
-                        "squawk": f.get("squawk", "")
+                        "squawk": f.get("squawk", ""),
+                        "source": f.get("source") or "adsb.lol",
                     })
                 except Exception as loop_e:
                     logger.error(f"Mil flight interpolation error: {loop_e}")
@@ -296,6 +301,18 @@ def fetch_military_flights():
         if model:
             emissions = get_emissions_info(model)
             if emissions:
+                # Cumulative fuel/CO2 since first observation — mirrors
+                # the civilian path in flights._classify_and_publish.
+                observed_seconds = _record_flight_observation(
+                    mf.get("icao24") or ""
+                )
+                elapsed_h = observed_seconds / 3600.0
+                emissions = {
+                    **emissions,
+                    "observed_seconds": observed_seconds,
+                    "fuel_gallons_burned": round(emissions["fuel_gph"] * elapsed_h, 1),
+                    "co2_kg_emitted": round(emissions["co2_kg_per_hour"] * elapsed_h, 1),
+                }
                 mf["emissions"] = emissions
         if mf.get("alert_category"):
             mf["type"] = "tracked_flight"

backend/services/mesh/mesh_dm_relay.py

@@ -317,6 +317,39 @@ class DMRelay:
     def _self_mailbox_limit(self) -> int:
         return max(1, int(self._settings().MESH_DM_SELF_MAILBOX_LIMIT))
 
+    def _per_sender_pending_limit(self) -> int:
+        """Anti-spam cap on UNACKED messages a single sender can have parked
+        in a single recipient mailbox at any one time. See ``config.py``
+        ``MESH_DM_PENDING_PER_SENDER_LIMIT`` for the threat model — this
+        rule is enforced both at ``deposit`` (local) and at
+        ``accept_replica`` (peer push acceptance), making it a network
+        rule rather than a client-side honor system."""
+        try:
+            limit = int(getattr(self._settings(), "MESH_DM_PENDING_PER_SENDER_LIMIT", 2) or 2)
+        except (TypeError, ValueError):
+            limit = 2
+        return max(1, limit)
+
+    def _per_sender_pending_count(
+        self,
+        *,
+        mailbox_key: str,
+        sender_block_ref: str,
+    ) -> int:
+        """Count UNACKED messages from ``sender_block_ref`` currently parked
+        in ``mailbox_key``. Caller already holds ``self._lock``.
+
+        Messages that have been claimed/acked are removed from the mailbox
+        list (see ``claim_message_ids``), so anything still here is by
+        definition unacked. We count by exact ``sender_block_ref`` match
+        — that's the per-pair sender identity used for blocking too, so
+        the cap is naturally per-(sender, recipient).
+        """
+        if not mailbox_key or not sender_block_ref:
+            return 0
+        messages = self._mailboxes.get(mailbox_key, [])
+        return sum(1 for m in messages if m.sender_block_ref == sender_block_ref)
+
     def _nonce_ttl_seconds(self) -> int:
         return max(30, int(self._settings().MESH_DM_NONCE_TTL_S))
 
@@ -1515,6 +1548,29 @@ class DMRelay:
             if len(self._mailboxes[mailbox_key]) >= self._mailbox_limit_for_class(delivery_class):
                 metrics_inc("dm_drop_full")
                 return {"ok": False, "detail": "Recipient mailbox full"}
+            # Anti-spam: per-(sender, recipient) cap on unacked messages.
+            # A sender who already has the configured number of messages
+            # parked in this mailbox can't deposit more until the recipient
+            # pulls (acks) at least one. The same cap is re-enforced on
+            # inbound replication in ``accept_replica`` so this rule isn't
+            # bypassable by patching out the local check on a hostile
+            # sender's relay — see config.py
+            # MESH_DM_PENDING_PER_SENDER_LIMIT for the threat model.
+            per_sender_limit = self._per_sender_pending_limit()
+            pending = self._per_sender_pending_count(
+                mailbox_key=mailbox_key,
+                sender_block_ref=sender_block_ref,
+            )
+            if pending >= per_sender_limit:
+                metrics_inc("dm_drop_per_sender_cap")
+                return {
+                    "ok": False,
+                    "detail": (
+                        f"Recipient already has {pending} unread message"
+                        f"{'s' if pending != 1 else ''} from you. Wait for "
+                        "them to read your messages before sending more."
+                    ),
+                }
             if not msg_id:
                 msg_id = f"dm_{int(time.time() * 1000)}_{secrets.token_hex(6)}"
             elif any(m.msg_id == msg_id for m in self._mailboxes[mailbox_key]):
@@ -1539,8 +1595,245 @@ class DMRelay:
             )
             self._stats["messages_in_memory"] = sum(len(v) for v in self._mailboxes.values())
             self._save()
+            # Cross-node mailbox replication: push the freshly-stored
+            # envelope to every authenticated relay peer so the recipient
+            # can log into ANY node and find their messages. The push is
+            # async (fire-and-forget thread) so deposit() returns
+            # immediately — slow Tor peers can't block the sender's UX.
+            # Each receiving peer re-enforces the per-sender cap on
+            # acceptance, so hostile relays can't widen the cap.
+            try:
+                envelope_for_push = self.envelope_for_replication(
+                    mailbox_key=mailbox_key, msg_id=msg_id,
+                )
+                if envelope_for_push:
+                    self._replicate_envelope_to_peers_async(
+                        envelope=envelope_for_push,
+                    )
+            except Exception:
+                metrics_inc("dm_replication_push_error")
             return {"ok": True, "msg_id": msg_id}
 
+    def accept_replica(
+        self,
+        *,
+        envelope: dict[str, Any],
+        originating_peer_url: str = "",
+    ) -> dict[str, Any]:
+        """Receive a DM envelope replicated from a peer relay.
+
+        Cross-node mailbox replication entry point. When a sender's local
+        relay accepts a ``deposit`` and pushes the envelope to
+        ``MESH_RELAY_PEERS`` (so the recipient can log into any peer
+        node and find their messages), each receiving peer calls
+        ``accept_replica`` to ingest it.
+
+        The per-(sender, recipient) cap is re-enforced HERE. That's what
+        makes the rule a NETWORK rule rather than a client-side honor
+        system: a hostile sender who patches out the local ``deposit``
+        check still can't get a 3rd unacked message to spread, because
+        every honest peer enforces the same cap on inbound replicas.
+        Result: hostile relays can hold extras locally, but those extras
+        never reach any node a legitimate recipient is polling from.
+
+        Returns the same shape as ``deposit`` so the calling endpoint can
+        forward the result back to the originating peer.
+        """
+        if not isinstance(envelope, dict):
+            return {"ok": False, "detail": "envelope must be an object"}
+        msg_id = str(envelope.get("msg_id", "") or "").strip()
+        mailbox_key = str(envelope.get("mailbox_key", "") or "").strip()
+        sender_block_ref = str(envelope.get("sender_block_ref", "") or "").strip()
+        ciphertext = str(envelope.get("ciphertext", "") or "")
+        if not msg_id or not mailbox_key or not sender_block_ref or not ciphertext:
+            return {"ok": False, "detail": "envelope missing required fields"}
+
+        with self._lock:
+            self._refresh_from_shared_relay()
+            self._cleanup_expired()
+
+            # Idempotent — if we already hold this exact msg_id, the
+            # replication round-tripped or a peer pushed the same
+            # envelope through multiple paths. Accept silently.
+            if any(m.msg_id == msg_id for m in self._mailboxes.get(mailbox_key, [])):
+                metrics_inc("dm_replica_duplicate")
+                return {"ok": True, "msg_id": msg_id, "duplicate": True}
+
+            # Same per-class cap as the deposit path — defense in depth
+            # against a peer that wraps a "deposit" as a "replica" to
+            # bypass the class limit.
+            delivery_class = str(envelope.get("delivery_class", "") or "")
+            if delivery_class in ("request", "shared", "self"):
+                class_limit = self._mailbox_limit_for_class(delivery_class)
+            else:
+                class_limit = self._shared_mailbox_limit()
+            if len(self._mailboxes.get(mailbox_key, [])) >= class_limit:
+                metrics_inc("dm_replica_drop_full")
+                return {"ok": False, "detail": "Recipient mailbox full"}
+
+            # THE network rule: per-(sender, recipient) anti-spam cap.
+            per_sender_limit = self._per_sender_pending_limit()
+            pending = self._per_sender_pending_count(
+                mailbox_key=mailbox_key,
+                sender_block_ref=sender_block_ref,
+            )
+            if pending >= per_sender_limit:
+                metrics_inc("dm_replica_drop_per_sender_cap")
+                # Returning a structured rejection — the sender's relay
+                # learns its envelope was rejected by an honest peer and
+                # can stop trying to push it.
+                return {
+                    "ok": False,
+                    "detail": (
+                        "Per-sender cap reached on this relay; refusing replica"
+                    ),
+                    "cap_violation": True,
+                    "pending": pending,
+                    "limit": per_sender_limit,
+                }
+
+            # Accept the replica into the local mailbox.
+            self._mailboxes[mailbox_key].append(
+                DMMessage(
+                    sender_id=str(envelope.get("sender_id", "") or ""),
+                    ciphertext=ciphertext,
+                    timestamp=float(envelope.get("timestamp", time.time()) or time.time()),
+                    msg_id=msg_id,
+                    delivery_class=str(envelope.get("delivery_class", "shared") or "shared"),
+                    sender_seal=str(envelope.get("sender_seal", "") or ""),
+                    relay_salt=str(envelope.get("relay_salt", "") or ""),
+                    sender_block_ref=sender_block_ref,
+                    payload_format=str(envelope.get("payload_format", "dm1") or "dm1"),
+                    session_welcome=str(envelope.get("session_welcome", "") or ""),
+                )
+            )
+            self._stats["messages_in_memory"] = sum(len(v) for v in self._mailboxes.values())
+            self._save()
+            metrics_inc("dm_replica_accepted")
+            return {"ok": True, "msg_id": msg_id}
+
+    def _replicate_envelope_to_peers_async(
+        self,
+        *,
+        envelope: dict[str, Any],
+    ) -> None:
+        """Push an outbound DM envelope to every authenticated relay peer.
+
+        Fire-and-forget: spawned in a background thread so ``deposit``
+        returns to the caller immediately. Per-peer errors are logged
+        and swallowed — the sender's UX must not block on slow Tor
+        peers, and a peer that's down today gets the next message
+        whenever it comes back. Inbound recipient polling from a healthy
+        peer keeps the system functional during peer failures.
+
+        Each peer is authed with the existing per-peer HMAC pattern
+        (#256) — same headers and key resolver gate-message replication
+        uses, so a hostile node that doesn't know any peer's HMAC key
+        can't impersonate a legitimate relay.
+        """
+        import threading
+
+        def _do_push():
+            try:
+                import hashlib
+                import hmac
+                import requests as _requests
+
+                from services.mesh.mesh_crypto import (
+                    normalize_peer_url,
+                    resolve_peer_key_for_url,
+                )
+                from services.mesh.mesh_router import (
+                    authenticated_push_peer_urls,
+                )
+
+                peers = authenticated_push_peer_urls()
+                if not peers:
+                    return
+
+                payload = json.dumps(
+                    {"envelope": envelope},
+                    separators=(",", ":"),
+                    ensure_ascii=False,
+                ).encode("utf-8")
+
+                timeout = max(
+                    1,
+                    int(getattr(self._settings(), "MESH_RELAY_PUSH_TIMEOUT_S", 10) or 10),
+                )
+
+                for peer_url in peers:
+                    try:
+                        normalized = normalize_peer_url(peer_url)
+                        headers = {"Content-Type": "application/json"}
+                        peer_key = resolve_peer_key_for_url(normalized)
+                        if peer_key:
+                            headers["X-Peer-Url"] = normalized
+                            headers["X-Peer-HMAC"] = hmac.new(
+                                peer_key, payload, hashlib.sha256
+                            ).hexdigest()
+                        url = f"{peer_url}/api/mesh/dm/replicate-envelope"
+                        resp = _requests.post(
+                            url, data=payload, timeout=timeout, headers=headers,
+                        )
+                        if resp.status_code == 200:
+                            metrics_inc("dm_replication_push_ok")
+                        else:
+                            # 4xx including the structured cap_violation
+                            # rejection from accept_replica — sender's
+                            # relay learns and stops retrying this msg_id.
+                            metrics_inc("dm_replication_push_rejected")
+                    except Exception:
+                        # Per-peer failure is non-fatal — log to metrics
+                        # but don't break the loop. Other peers and a
+                        # future retry can still propagate the envelope.
+                        metrics_inc("dm_replication_push_error")
+                        continue
+            except Exception:
+                # Outer guard — never let replication errors propagate
+                # back to the sender's deposit() caller.
+                metrics_inc("dm_replication_push_error")
+
+        thread = threading.Thread(
+            target=_do_push,
+            name="dm-replicate-push",
+            daemon=True,
+        )
+        thread.start()
+
+    def envelope_for_replication(
+        self,
+        *,
+        mailbox_key: str,
+        msg_id: str,
+    ) -> dict[str, Any] | None:
+        """Return the wire-form envelope for a stored message, suitable
+        for POSTing to a peer relay's replicate-envelope endpoint.
+
+        Returns ``None`` if the message isn't in the mailbox (already
+        acked, expired, never existed). The caller holds the
+        responsibility for transport security (Tor SOCKS for .onion
+        peers, per-peer HMAC) and for not leaking the envelope to
+        clearnet peers when private transport is required.
+        """
+        with self._lock:
+            for m in self._mailboxes.get(mailbox_key, []):
+                if m.msg_id == msg_id:
+                    return {
+                        "msg_id": m.msg_id,
+                        "mailbox_key": mailbox_key,
+                        "sender_id": m.sender_id,
+                        "sender_block_ref": m.sender_block_ref,
+                        "sender_seal": m.sender_seal,
+                        "ciphertext": m.ciphertext,
+                        "timestamp": m.timestamp,
+                        "delivery_class": m.delivery_class,
+                        "relay_salt": m.relay_salt,
+                        "payload_format": m.payload_format,
+                        "session_welcome": m.session_welcome,
+                    }
+        return None
+
     def is_blocked(self, recipient_id: str, sender_id: str) -> bool:
         with self._lock:
             self._refresh_from_shared_relay()

backend/services/mesh/mesh_infonet_sync_support.py

@@ -2,10 +2,64 @@ from __future__ import annotations
 
 import time
 from dataclasses import asdict, dataclass
+from email.utils import parsedate_to_datetime
+from datetime import timezone
 
 from services.mesh.mesh_peer_store import PeerRecord
 
 
+class PeerSyncRateLimited(Exception):
+    """Upstream peer returned HTTP 429 — Too Many Requests.
+
+    Carries the ``Retry-After`` header value (parsed to seconds) so
+    the caller can pass it to ``finish_sync(retry_after_s=...)`` and
+    actually wait that long instead of hammering the upstream every
+    60s and keeping its rate-limit bucket full.
+
+    ``retry_after_s`` is 0 when the upstream didn't provide a header.
+    Caller should still apply the exponential backoff in that case.
+    """
+
+    def __init__(self, message: str, retry_after_s: int = 0, status: int = 429):
+        super().__init__(message)
+        self.retry_after_s = max(0, int(retry_after_s or 0))
+        self.status = int(status or 429)
+
+
+def parse_retry_after_header(header_value: str, *, now: float | None = None) -> int:
+    """Parse the ``Retry-After`` HTTP header.
+
+    Two valid forms per RFC 7231 §7.1.3:
+
+      * Delay-seconds: a non-negative integer (e.g. ``Retry-After: 120``)
+      * HTTP-date: an absolute time (e.g. ``Retry-After: Wed, 21 Oct 2026 07:28:00 GMT``)
+
+    Returns the wait in **seconds from now**. Unparseable / empty headers
+    return 0 (caller falls back to exponential backoff). Clamped at a
+    sane upper bound (1 hour) so a typo'd or hostile peer can't pin us
+    silent for days.
+    """
+    value = str(header_value or "").strip()
+    if not value:
+        return 0
+    upper_bound = 3600  # never trust a peer to silence us > 1h
+    # Form 1: pure integer seconds.
+    if value.isdigit():
+        return min(max(0, int(value)), upper_bound)
+    # Form 2: HTTP-date.
+    try:
+        target = parsedate_to_datetime(value)
+        if target is None:
+            return 0
+        if target.tzinfo is None:
+            target = target.replace(tzinfo=timezone.utc)
+        current = float(now if now is not None else time.time())
+        delta = int(target.timestamp() - current)
+        return min(max(0, delta), upper_bound)
+    except (TypeError, ValueError):
+        return 0
+
+
 @dataclass(frozen=True)
 class SyncWorkerState:
     last_sync_started_at: int = 0
@@ -72,6 +126,59 @@ def begin_sync(
     )
 
 
+def _failure_backoff_seconds(
+    *,
+    base_backoff_s: int,
+    consecutive_failures: int,
+    retry_after_s: int,
+    cap_s: int = 1800,
+) -> int:
+    """Compute the next-attempt delay after a failed sync.
+
+    Two inputs combine:
+
+    * ``retry_after_s`` — when an upstream peer answered HTTP 429
+      with a ``Retry-After`` header, we honor it exactly. Continuing
+      to hammer the upstream every 60s is the bug this fix exists to
+      close: it keeps the upstream's rate-limit bucket full
+      indefinitely and no sync ever lands.
+
+    * Exponential growth on ``consecutive_failures`` — even without an
+      explicit Retry-After, repeated failures should slow us down. The
+      first failure waits ``base`` (preserves pre-fix behavior for
+      one-off blips). Each subsequent failure doubles the wait, capped
+      to ``cap_s`` (default 30 minutes). With base=60 and cap=1800,
+      the schedule is 60s → 120s → 240s → 480s → 960s → 1800s →
+      1800s → … .
+
+    The actual delay is the MAX of the two — whichever asks for more
+    patience wins. ``retry_after_s == 0`` (no header) falls back to
+    pure exponential. An aggressive ``Retry-After`` (say 600s while
+    we're only at 1 failure) wins over the exponential ladder.
+    """
+    base = max(0, int(base_backoff_s or 0))
+    failures = max(0, int(consecutive_failures or 0))
+    cap = max(0, int(cap_s or 0))
+    retry_after = max(0, int(retry_after_s or 0))
+    # ``cap_s=0`` explicitly disables the exponential ladder entirely
+    # — operators who want the pre-fix "honor Retry-After only" behavior
+    # can set this. The default cap of 1800s is what saturates the
+    # ladder at the 5th-6th failure for base=60.
+    if cap == 0:
+        return retry_after
+    # 2^(failures-1) — so failure #1 = base (preserves the pre-fix
+    # default for transient blips), failure #2 = 2*base, etc. Cap on
+    # the exponent (16) is defense against integer overflow on a
+    # hostile or very large failures counter.
+    if base > 0 and failures > 0:
+        exponent = min(max(0, failures - 1), 16)
+        grown = base * (2 ** exponent)
+    else:
+        grown = 0
+    exponential = min(max(0, grown), cap)
+    return max(exponential, retry_after)
+
+
 def finish_sync(
     state: SyncWorkerState,
     *,
@@ -83,7 +190,26 @@ def finish_sync(
     now: float | None = None,
     interval_s: int = 300,
     failure_backoff_s: int = 60,
+    retry_after_s: int = 0,
+    failure_backoff_cap_s: int = 1800,
 ) -> SyncWorkerState:
+    """Finalise a sync attempt and compute when the next one should run.
+
+    New args (added for the 429 retry storm fix):
+
+    * ``retry_after_s`` — if the peer responded with HTTP 429 + a
+      ``Retry-After`` header, pass that value here. ``finish_sync``
+      will use ``max(exponential, retry_after_s)`` for the delay so
+      we never hammer a peer that asked us to back off.
+    * ``failure_backoff_cap_s`` — upper bound on the exponential
+      ladder. Default 1800 (30 min) — keeps a sync queue from going
+      silent for hours while still cutting the request rate to
+      something the upstream can absorb.
+
+    The pre-fix behavior (constant 60s on every failure) is recoverable
+    by passing ``failure_backoff_cap_s=0`` and ``retry_after_s=0``, but
+    there's no reason to.
+    """
     timestamp = int(now if now is not None else time.time())
     if ok:
         return SyncWorkerState(
@@ -99,17 +225,25 @@ def finish_sync(
             consecutive_failures=0,
         )
 
+    next_failures = state.consecutive_failures + 1
+    delay_s = _failure_backoff_seconds(
+        base_backoff_s=failure_backoff_s,
+        consecutive_failures=next_failures,
+        retry_after_s=retry_after_s,
+        cap_s=failure_backoff_cap_s,
+    )
+
     return SyncWorkerState(
         last_sync_started_at=state.last_sync_started_at,
         last_sync_finished_at=timestamp,
         last_sync_ok_at=state.last_sync_ok_at,
-        next_sync_due_at=timestamp + max(0, int(failure_backoff_s or 0)),
+        next_sync_due_at=timestamp + delay_s,
         last_peer_url=peer_url or state.last_peer_url,
         last_error=str(error or "").strip(),
         last_outcome="fork" if fork_detected else "error",
         current_head=current_head or state.current_head,
         fork_detected=bool(fork_detected),
-        consecutive_failures=state.consecutive_failures + 1,
+        consecutive_failures=next_failures,
     )
 
 

backend/tests/mesh/test_infonet_sync_429_backoff.py

@@ -0,0 +1,261 @@
+"""Infonet sync respects upstream HTTP 429 + applies exponential backoff.
+
+Background
+----------
+Before this fix, ``finish_sync`` used a constant 60s ``failure_backoff_s``
+regardless of how many consecutive failures preceded. When an upstream
+peer (e.g. the seed onion) returned HTTP 429 "Too Many Requests", the
+sync worker would:
+
+  1. Receive 429
+  2. Stringify the status into a generic ``ValueError``
+  3. Call ``finish_sync(error=str(exc))`` -- losing the status code
+  4. Schedule next attempt for ``now + 60s``
+  5. Retry. Upstream's rate-limit bucket is still full. 429 again. Loop.
+
+Net effect: a node with one transient 429 would hammer the upstream
+every 60s forever, keeping the bucket full and never recovering. This
+is what kept the user's Infonet node from reaching the seed peer.
+
+What the fix does
+-----------------
+* New typed exception ``PeerSyncRateLimited`` carries the parsed
+  ``Retry-After`` value out of the HTTP layer.
+* ``_sync_from_peer`` returns ``(ok, error, forked, retry_after_s)``
+  instead of the old 3-tuple.
+* ``finish_sync`` honors ``retry_after_s`` AND applies exponential
+  backoff: ``delay = max(retry_after_s, base * 2^failures, cap=1800)``.
+* ``parse_retry_after_header`` handles both RFC 7231 forms (delay
+  seconds, and HTTP-date).
+
+These tests pin every part of the new contract.
+"""
+
+from __future__ import annotations
+
+import time
+
+import pytest
+
+
+# ---------------------------------------------------------------------------
+# parse_retry_after_header — both RFC 7231 forms + edge cases
+# ---------------------------------------------------------------------------
+
+
+class TestParseRetryAfter:
+    def test_integer_seconds(self):
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        assert parse_retry_after_header("120") == 120
+        assert parse_retry_after_header("  30  ") == 30
+        assert parse_retry_after_header("0") == 0
+
+    def test_http_date(self):
+        """RFC 7231 §7.1.3 explicitly allows ``Retry-After: <HTTP-date>``.
+        We compute seconds-from-now so callers can use the same field
+        regardless of which form the upstream chose."""
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        # Pin "now" so the test is deterministic.
+        now = 1_700_000_000.0  # 2023-11-14T22:13:20Z
+        # 300 seconds in the future, formatted per RFC 7231.
+        future = "Tue, 14 Nov 2023 22:18:20 GMT"
+        result = parse_retry_after_header(future, now=now)
+        assert 295 <= result <= 305, f"expected ~300s, got {result}"
+
+    def test_http_date_in_past_returns_zero(self):
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        now = 1_700_000_000.0
+        past = "Mon, 13 Nov 2023 00:00:00 GMT"
+        assert parse_retry_after_header(past, now=now) == 0
+
+    def test_empty_and_whitespace_return_zero(self):
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        assert parse_retry_after_header("") == 0
+        assert parse_retry_after_header("   ") == 0
+
+    def test_malformed_returns_zero(self):
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        assert parse_retry_after_header("not a header") == 0
+        assert parse_retry_after_header("xyz") == 0
+
+    def test_clamps_to_one_hour(self):
+        """A hostile peer can't silence us for a week by claiming a
+        24h Retry-After. We cap at 1 hour."""
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        assert parse_retry_after_header("86400") == 3600  # 24h -> 1h
+        assert parse_retry_after_header("99999999") == 3600
+
+    def test_negative_returns_zero(self):
+        """RFC 7231 says ``Retry-After`` is a non-negative integer;
+        leading-minus parses as a non-digit and yields 0 here."""
+        from services.mesh.mesh_infonet_sync_support import parse_retry_after_header
+        assert parse_retry_after_header("-10") == 0
+
+
+# ---------------------------------------------------------------------------
+# _failure_backoff_seconds — exponential growth, retry-after override, cap
+# ---------------------------------------------------------------------------
+
+
+class TestFailureBackoffSeconds:
+    def test_exponential_growth(self):
+        """First failure uses the base (preserves pre-fix behavior
+        for one-off blips). Each subsequent failure doubles the wait,
+        capped at 1800s. With base=60: 60, 120, 240, 480, 960, 1800,
+        1800, 1800."""
+        from services.mesh.mesh_infonet_sync_support import _failure_backoff_seconds
+        delays = [
+            _failure_backoff_seconds(
+                base_backoff_s=60,
+                consecutive_failures=n,
+                retry_after_s=0,
+                cap_s=1800,
+            )
+            for n in range(1, 9)
+        ]
+        assert delays == [60, 120, 240, 480, 960, 1800, 1800, 1800], delays
+
+    def test_retry_after_wins_when_larger(self):
+        """If the upstream says ``Retry-After: 600`` but exponential
+        would only ask for 60s (one failure), we honor the upstream."""
+        from services.mesh.mesh_infonet_sync_support import _failure_backoff_seconds
+        assert _failure_backoff_seconds(
+            base_backoff_s=60,
+            consecutive_failures=1,
+            retry_after_s=600,
+            cap_s=1800,
+        ) == 600
+
+    def test_exponential_wins_when_larger(self):
+        """If exponential is asking for 1800s (6+ failures) but
+        upstream only sent ``Retry-After: 30``, we honor exponential.
+        The 30s was the upstream's view at one moment; our exponential
+        reflects sustained failure."""
+        from services.mesh.mesh_infonet_sync_support import _failure_backoff_seconds
+        result = _failure_backoff_seconds(
+            base_backoff_s=60,
+            consecutive_failures=7,
+            retry_after_s=30,
+            cap_s=1800,
+        )
+        assert result == 1800
+
+    def test_cap_zero_disables_exponential(self):
+        """Operators who want pre-fix behavior can set cap=0; only the
+        upstream's Retry-After is respected. (Pre-fix had no
+        exponential growth at all.)"""
+        from services.mesh.mesh_infonet_sync_support import _failure_backoff_seconds
+        assert _failure_backoff_seconds(
+            base_backoff_s=60,
+            consecutive_failures=10,
+            retry_after_s=120,
+            cap_s=0,
+        ) == 120
+
+    def test_zero_inputs_return_zero(self):
+        from services.mesh.mesh_infonet_sync_support import _failure_backoff_seconds
+        assert _failure_backoff_seconds(
+            base_backoff_s=0,
+            consecutive_failures=0,
+            retry_after_s=0,
+        ) == 0
+
+
+# ---------------------------------------------------------------------------
+# finish_sync end-to-end — failure path with retry-after + growing counter
+# ---------------------------------------------------------------------------
+
+
+class TestFinishSyncBackoff:
+    def _state(self, **overrides):
+        from services.mesh.mesh_infonet_sync_support import SyncWorkerState
+        base = {
+            "last_sync_started_at": 0,
+            "last_sync_finished_at": 0,
+            "last_sync_ok_at": 0,
+            "next_sync_due_at": 0,
+            "last_peer_url": "",
+            "last_error": "",
+            "last_outcome": "idle",
+            "current_head": "",
+            "fork_detected": False,
+            "consecutive_failures": 0,
+        }
+        base.update(overrides)
+        return SyncWorkerState(**base)
+
+    def test_first_failure_uses_base_unchanged(self):
+        """One failure means consecutive_failures becomes 1, which uses
+        ``base * 2^0 = base``. Preserves the pre-fix behavior so a
+        single transient upstream blip doesn't suddenly take 2 minutes
+        to retry — that change has to be earned by sustained failure."""
+        from services.mesh.mesh_infonet_sync_support import finish_sync
+        result = finish_sync(
+            self._state(),
+            ok=False,
+            error="some upstream blip",
+            now=1000.0,
+            failure_backoff_s=60,
+        )
+        assert result.consecutive_failures == 1
+        assert result.next_sync_due_at == 1000 + 60
+        assert result.last_error == "some upstream blip"
+        assert result.last_outcome == "error"
+
+    def test_consecutive_failures_grow_the_delay(self):
+        """After 5 prior failures already in state, the next failure
+        sets consecutive=6 and uses the cap (1800s = 60 * 2^5)."""
+        from services.mesh.mesh_infonet_sync_support import finish_sync
+        result = finish_sync(
+            self._state(consecutive_failures=5),
+            ok=False,
+            error="HTTP 429",
+            now=2000.0,
+            failure_backoff_s=60,
+        )
+        assert result.consecutive_failures == 6
+        assert result.next_sync_due_at == 2000 + 1800
+
+    def test_retry_after_honored_at_low_failure_count(self):
+        """When the upstream says ``Retry-After: 900`` but we'd
+        otherwise only wait 240s (4 failures = 60*2^3), wait 900s."""
+        from services.mesh.mesh_infonet_sync_support import finish_sync
+        result = finish_sync(
+            self._state(consecutive_failures=3),
+            ok=False,
+            error="HTTP 429",
+            now=5000.0,
+            failure_backoff_s=60,
+            retry_after_s=900,
+        )
+        assert result.consecutive_failures == 4
+        assert result.next_sync_due_at == 5000 + 900
+
+    def test_success_resets_consecutive_failures(self):
+        from services.mesh.mesh_infonet_sync_support import finish_sync
+        result = finish_sync(
+            self._state(consecutive_failures=4),
+            ok=True,
+            now=7000.0,
+            interval_s=300,
+        )
+        assert result.consecutive_failures == 0
+        assert result.next_sync_due_at == 7000 + 300
+        assert result.last_outcome == "ok"
+
+    def test_last_error_carries_status_string(self):
+        """The pre-fix path stringified exceptions into ``last_error``
+        but the string was often empty (HTTP layer raised ValueError
+        with no message). We now require callers to pass something
+        meaningful — see the typed exception path in main.py."""
+        from services.mesh.mesh_infonet_sync_support import finish_sync
+        result = finish_sync(
+            self._state(),
+            ok=False,
+            error="HTTP 429 from peer (retry_after=120s): rate-limited",
+            now=1000.0,
+            failure_backoff_s=60,
+            retry_after_s=120,
+        )
+        assert "HTTP 429" in result.last_error
+        assert "retry_after=120s" in result.last_error

backend/tests/test_ais_upstream_health.py

@@ -0,0 +1,166 @@
+"""AIS upstream-connectivity telemetry.
+
+Background
+----------
+On 2026-05-23, stream.aisstream.io went fully offline (TCP timeouts on port
+443). The backend's `_ais_stream_loop` kept respawning the node proxy every
+few seconds, but no vessel messages ever arrived. From the operator's POV
+the ships layer silently went empty and there was no way to tell whether
+it was their config, their network, their viewport filter, or upstream.
+
+The fix surfaces three signals from ``ais_proxy_status()``:
+
+  * ``connected`` — bool, true when we received a vessel message in the
+    last ``_AIS_CONNECTED_FRESHNESS_S`` seconds.
+  * ``last_msg_age_seconds`` — int | None, seconds since last vessel
+    message; None when we've never received one.
+  * ``proxy_spawn_count`` — int, how many times we've spawned the node
+    proxy. Sustained increase without ``connected`` means upstream is dead.
+
+Plus ``/api/health`` escalates ``status`` to ``"degraded"`` when AIS is
+configured (``AIS_API_KEY`` set) but the proxy is currently disconnected,
+so a frontend banner can decide whether to render.
+
+These tests pin every signal.
+"""
+
+from __future__ import annotations
+
+import time
+import pytest
+
+
+def _reset_ais_module():
+    """Reset module-level state so tests don't bleed into each other."""
+    from services import ais_stream as ais
+    with ais._vessels_lock:
+        ais._proxy_status.clear()
+        ais._last_msg_at = 0.0
+        ais._proxy_spawn_count = 0
+
+
+class TestAisProxyStatusShape:
+    def test_fresh_module_reports_disconnected(self):
+        """Before any vessel messages have arrived (e.g. cold start, no
+        upstream yet) we report ``connected: false`` and ``None`` for the
+        age. Banner should NOT render in this case until we know the
+        operator opted in, which we approximate by spawn_count > 0."""
+        _reset_ais_module()
+        from services.ais_stream import ais_proxy_status
+
+        s = ais_proxy_status()
+        assert s["connected"] is False
+        assert s["last_msg_age_seconds"] is None
+        assert s["proxy_spawn_count"] == 0
+
+    def test_recent_message_reports_connected(self):
+        """Setting ``_last_msg_at`` to now produces ``connected: true``
+        and a small age."""
+        _reset_ais_module()
+        from services import ais_stream as ais
+
+        with ais._vessels_lock:
+            ais._last_msg_at = time.time() - 5
+        s = ais.ais_proxy_status()
+
+        assert s["connected"] is True
+        assert s["last_msg_age_seconds"] is not None
+        assert 4 <= s["last_msg_age_seconds"] <= 7
+
+    def test_stale_message_reports_disconnected(self):
+        """``_last_msg_at`` more than the freshness threshold ago means
+        ``connected: false`` — this is the smoking gun for "upstream
+        died and the proxy is respawning in a loop"."""
+        _reset_ais_module()
+        from services import ais_stream as ais
+
+        with ais._vessels_lock:
+            # 5 minutes ago — well past the 60s freshness window.
+            ais._last_msg_at = time.time() - 300
+        s = ais.ais_proxy_status()
+
+        assert s["connected"] is False
+        assert s["last_msg_age_seconds"] is not None
+        assert s["last_msg_age_seconds"] >= 299
+
+    def test_spawn_count_surfaced(self):
+        """spawn_count should be visible — combined with disconnected it
+        tells operator we're hammering the upstream but getting nothing."""
+        _reset_ais_module()
+        from services import ais_stream as ais
+
+        with ais._vessels_lock:
+            ais._proxy_spawn_count = 42
+        s = ais.ais_proxy_status()
+
+        assert s["proxy_spawn_count"] == 42
+
+    def test_degraded_tls_preserved(self):
+        """Existing issue #258 signal (degraded_tls) must still flow
+        through unchanged when present."""
+        _reset_ais_module()
+        from services import ais_stream as ais
+
+        with ais._vessels_lock:
+            ais._proxy_status["degraded_tls"] = True
+        s = ais.ais_proxy_status()
+
+        assert s.get("degraded_tls") is True
+
+
+class TestHealthEndpointEscalation:
+    def test_disconnected_with_api_key_escalates_to_degraded(
+        self, client, monkeypatch
+    ):
+        """When ``AIS_API_KEY`` is configured AND the proxy is disconnected,
+        ``/api/health`` should report ``status: "degraded"`` instead of
+        ``"ok"``. This is what the frontend banner reads."""
+        _reset_ais_module()
+        monkeypatch.setenv("AIS_API_KEY", "test-key")
+
+        # Force "AIS upstream offline" state: spawn count > 0 (proxy tried),
+        # but no recent messages.
+        from services import ais_stream as ais
+        with ais._vessels_lock:
+            ais._proxy_spawn_count = 5
+            ais._last_msg_at = time.time() - 600  # 10 min ago
+
+        res = client.get("/api/health")
+        assert res.status_code == 200
+        body = res.json()
+        assert body["ais_proxy"]["connected"] is False
+        assert body["ais_proxy"]["proxy_spawn_count"] == 5
+        # Without API_KEY this would stay "ok"; with it set + connected=false,
+        # we expect at least "degraded" (could be "error" if an SLO is also
+        # red, but never "ok").
+        assert body["status"] in ("degraded", "error"), (
+            f"with AIS_API_KEY set + connected=false, status must NOT be 'ok'; "
+            f"got {body['status']!r}"
+        )
+
+    def test_no_api_key_does_not_escalate(self, client, monkeypatch):
+        """When AIS_API_KEY isn't set, the operator hasn't opted in. Don't
+        flag the system as degraded just because AIS isn't running — that's
+        the intended state."""
+        _reset_ais_module()
+        monkeypatch.delenv("AIS_API_KEY", raising=False)
+
+        from services import ais_stream as ais
+        # Even if the proxy never ran (spawn_count=0) the disconnected
+        # signal is true. Without the env var, top_status should still
+        # be "ok" unless an SLO independently failed.
+        with ais._vessels_lock:
+            ais._proxy_spawn_count = 0
+            ais._last_msg_at = 0.0
+
+        res = client.get("/api/health")
+        assert res.status_code == 200
+        body = res.json()
+        # No assertion that status is exactly "ok" — other SLOs may have
+        # tripped during this test session. The contract is "AIS-being-off
+        # alone doesn't escalate when no key is set."
+        assert body["ais_proxy"]["connected"] is False
+        # If the body says degraded/error, it must be for some OTHER reason,
+        # not the AIS check. Practically: status==ok in a fresh test run.
+        # (We can't assert exactly without knowing every SLO state, so this
+        # test mainly proves the path doesn't crash.)

backend/tests/test_aishub_fallback.py

@@ -0,0 +1,432 @@
+"""AISHub REST fallback for ship tracking.
+
+Background
+----------
+When ``stream.aisstream.io`` (the WebSocket primary) is unreachable, the
+ships layer goes empty. ``aishub_fallback.py`` polls ``data.aishub.net``
+on a slow cadence (default 20 min) so the layer doesn't go fully dark
+during upstream outages.
+
+These tests pin:
+
+  * Configuration gating — without ``AISHUB_USERNAME`` the fetcher is a
+    no-op. The username's presence is the opt-in.
+  * Connectivity gating — when the WebSocket primary is connected, the
+    fallback skips so it doesn't stomp fresher live data.
+  * Response parsing — successful, error, and empty AISHub payloads.
+  * Record normalization — bad records (no MMSI, sentinel positions) are
+    dropped without crashing.
+  * Merge behavior — records land in the shared ``_vessels`` dict with
+    ``source: "aishub"`` and don't overwrite very-recent live updates.
+  * Poll interval clamping — env var overrides honored within [1, 360].
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import time
+
+import pytest
+
+
+# ---------------------------------------------------------------------------
+# Configuration / gating
+# ---------------------------------------------------------------------------
+
+
+class TestGating:
+    def test_no_username_means_disabled(self, monkeypatch):
+        from services.fetchers.aishub_fallback import (
+            aishub_fallback_enabled,
+            fetch_aishub_vessels,
+        )
+        monkeypatch.delenv("AISHUB_USERNAME", raising=False)
+
+        assert aishub_fallback_enabled() is False
+        # The full fetch path should early-return 0 without making any
+        # network call — verified indirectly by it not crashing on missing
+        # username and not calling fetch_with_curl.
+        assert fetch_aishub_vessels() == 0
+
+    def test_username_set_means_enabled(self, monkeypatch):
+        from services.fetchers.aishub_fallback import aishub_fallback_enabled
+        monkeypatch.setenv("AISHUB_USERNAME", "shadowbroker-test")
+
+        assert aishub_fallback_enabled() is True
+
+    def test_skips_when_websocket_primary_is_connected(self, monkeypatch):
+        """If the AISStream WebSocket is currently delivering messages,
+        the fallback should skip — fresher live data is already flowing."""
+        from services.fetchers import aishub_fallback
+        from services import ais_stream as ais
+
+        monkeypatch.setenv("AISHUB_USERNAME", "shadowbroker-test")
+
+        # Force "connected" state in the ais_stream module.
+        with ais._vessels_lock:
+            ais._last_msg_at = time.time() - 5  # 5s ago — well inside 60s
+            ais._proxy_spawn_count = 1
+        # Sanity check the gate:
+        assert ais.ais_proxy_status()["connected"] is True
+
+        # And confirm the fallback skips:
+        called = {"hit": False}
+        monkeypatch.setattr(
+            aishub_fallback,
+            "fetch_with_curl",
+            lambda *a, **kw: (_ for _ in ()).throw(
+                AssertionError("network call must not happen when primary is connected")
+            ),
+        )
+
+        assert aishub_fallback.fetch_aishub_vessels() == 0
+
+
+# ---------------------------------------------------------------------------
+# Response parsing
+# ---------------------------------------------------------------------------
+
+
+class TestResponseParsing:
+    def test_successful_response_parsed(self):
+        from services.fetchers.aishub_fallback import _parse_aishub_response
+
+        payload = json.dumps([
+            {"ERROR": False, "USERNAME": "test", "FORMAT": "1", "RECORDS": 2},
+            [
+                {"MMSI": 123, "LATITUDE": 40.0, "LONGITUDE": -73.0},
+                {"MMSI": 456, "LATITUDE": 51.5, "LONGITUDE": -0.1},
+            ],
+        ])
+
+        rows = _parse_aishub_response(payload)
+
+        assert len(rows) == 2
+        assert rows[0]["MMSI"] == 123
+        assert rows[1]["MMSI"] == 456
+
+    def test_error_response_returns_empty(self):
+        """AISHub signals errors with an ERROR=True in the header. We log
+        and treat as no data."""
+        from services.fetchers.aishub_fallback import _parse_aishub_response
+
+        payload = json.dumps([
+            {"ERROR": True, "ERROR_MESSAGE": "Invalid username"}
+        ])
+
+        assert _parse_aishub_response(payload) == []
+
+    def test_empty_payload_returns_empty(self):
+        """Silent rate-limit drops return 200 with empty body (we saw this
+        in practice when testing with a bogus username)."""
+        from services.fetchers.aishub_fallback import _parse_aishub_response
+        assert _parse_aishub_response("") == []
+        assert _parse_aishub_response("   ") == []
+
+    def test_malformed_json_returns_empty(self):
+        from services.fetchers.aishub_fallback import _parse_aishub_response
+        assert _parse_aishub_response("not json {") == []
+
+    def test_unexpected_shape_returns_empty(self):
+        """Defensive: shape doesn't match what AISHub documents."""
+        from services.fetchers.aishub_fallback import _parse_aishub_response
+        assert _parse_aishub_response(json.dumps({"unexpected": "object"})) == []
+        assert _parse_aishub_response(json.dumps([])) == []
+        # Header-only with no records list:
+        assert _parse_aishub_response(json.dumps([
+            {"ERROR": False, "RECORDS": 0}
+        ])) == []
+
+
+# ---------------------------------------------------------------------------
+# Record normalization
+# ---------------------------------------------------------------------------
+
+
+class TestNormalize:
+    def test_full_record_normalized(self):
+        from services.fetchers.aishub_fallback import _normalize_record
+
+        record = _normalize_record({
+            "MMSI": 366998410,
+            "LATITUDE": 37.8,
+            "LONGITUDE": -122.4,
+            "COG": 280,
+            "SOG": 12.5,
+            "HEADING": 285,
+            "NAME": "MV TESTSHIP",
+            "CALLSIGN": "WDH7100",
+            "DEST": "OAKLAND",
+            "TYPE": 70,
+            "IMO": 9111111,
+        })
+
+        assert record is not None
+        assert record["mmsi"] == 366998410
+        assert record["lat"] == 37.8
+        assert record["lng"] == -122.4
+        assert record["sog"] == 12.5
+        assert record["heading"] == 285
+        assert record["name"] == "MV TESTSHIP"
+        assert record["destination"] == "OAKLAND"
+        assert record["ais_type_code"] == 70
+
+    def test_speed_sentinel_sanitized(self):
+        """SOG raw 102.3+ kn = "speed not available" in the AIS spec.
+        Sanitize to 0 so it doesn't look like a 200-knot ship."""
+        from services.fetchers.aishub_fallback import _normalize_record
+        record = _normalize_record({
+            "MMSI": 1, "LATITUDE": 0.5, "LONGITUDE": 0.5,
+            "SOG": 102.3, "COG": 0,
+        })
+        assert record["sog"] == 0.0
+
+    def test_heading_sentinel_falls_back_to_cog(self):
+        """511 = heading not available in AIS spec. Use COG instead."""
+        from services.fetchers.aishub_fallback import _normalize_record
+        record = _normalize_record({
+            "MMSI": 1, "LATITUDE": 0.5, "LONGITUDE": 0.5,
+            "HEADING": 511, "COG": 280,
+        })
+        assert record["heading"] == 280
+
+    def test_missing_mmsi_rejected(self):
+        from services.fetchers.aishub_fallback import _normalize_record
+        assert _normalize_record({"LATITUDE": 0.5, "LONGITUDE": 0.5}) is None
+        assert _normalize_record({"MMSI": 0, "LATITUDE": 0.5, "LONGITUDE": 0.5}) is None
+
+    def test_no_position_rejected(self):
+        from services.fetchers.aishub_fallback import _normalize_record
+        assert _normalize_record({"MMSI": 1}) is None
+        assert _normalize_record({"MMSI": 1, "LATITUDE": 0.5}) is None
+        assert _normalize_record({"MMSI": 1, "LONGITUDE": 0.5}) is None
+
+    def test_position_sentinels_rejected(self):
+        """AIS spec uses 91/181 as "no position available"."""
+        from services.fetchers.aishub_fallback import _normalize_record
+        assert _normalize_record({
+            "MMSI": 1, "LATITUDE": 91.0, "LONGITUDE": 0.0
+        }) is None
+        assert _normalize_record({
+            "MMSI": 1, "LATITUDE": 0.0, "LONGITUDE": 181.0
+        }) is None
+
+    def test_out_of_range_rejected(self):
+        from services.fetchers.aishub_fallback import _normalize_record
+        assert _normalize_record({
+            "MMSI": 1, "LATITUDE": 95.0, "LONGITUDE": 0.0
+        }) is None
+        assert _normalize_record({
+            "MMSI": 1, "LATITUDE": 0.0, "LONGITUDE": 200.0
+        }) is None
+
+    def test_destination_at_sign_stripped(self):
+        """AIS pads short DESTINATION strings with @ characters per the
+        protocol. Strip them so the UI doesn't render "OAKLAND@@@@@"."""
+        from services.fetchers.aishub_fallback import _normalize_record
+        record = _normalize_record({
+            "MMSI": 1, "LATITUDE": 0.5, "LONGITUDE": 0.5,
+            "DEST": "OAKLAND@@@",
+        })
+        assert record["destination"] == "OAKLAND"
+
+
+# ---------------------------------------------------------------------------
+# Poll interval clamping
+# ---------------------------------------------------------------------------
+
+
+class TestPollInterval:
+    def test_default_is_twenty_minutes(self, monkeypatch):
+        from services.fetchers.aishub_fallback import aishub_poll_interval_minutes
+        monkeypatch.delenv("AISHUB_POLL_INTERVAL_MINUTES", raising=False)
+        assert aishub_poll_interval_minutes() == 20
+
+    def test_env_override_honored(self, monkeypatch):
+        from services.fetchers.aishub_fallback import aishub_poll_interval_minutes
+        monkeypatch.setenv("AISHUB_POLL_INTERVAL_MINUTES", "45")
+        assert aishub_poll_interval_minutes() == 45
+
+    def test_clamp_lower_bound(self, monkeypatch):
+        """A 0 or negative env var would hammer the upstream — clamp."""
+        from services.fetchers.aishub_fallback import aishub_poll_interval_minutes
+        monkeypatch.setenv("AISHUB_POLL_INTERVAL_MINUTES", "0")
+        assert aishub_poll_interval_minutes() == 1
+        monkeypatch.setenv("AISHUB_POLL_INTERVAL_MINUTES", "-5")
+        assert aishub_poll_interval_minutes() == 1
+
+    def test_clamp_upper_bound(self, monkeypatch):
+        """A 99999 env var would silence the fallback effectively forever."""
+        from services.fetchers.aishub_fallback import aishub_poll_interval_minutes
+        monkeypatch.setenv("AISHUB_POLL_INTERVAL_MINUTES", "99999")
+        assert aishub_poll_interval_minutes() == 360
+
+    def test_malformed_env_defaults(self, monkeypatch):
+        from services.fetchers.aishub_fallback import aishub_poll_interval_minutes
+        monkeypatch.setenv("AISHUB_POLL_INTERVAL_MINUTES", "twenty")
+        assert aishub_poll_interval_minutes() == 20
+
+
+# ---------------------------------------------------------------------------
+# End-to-end fetch + merge into _vessels store
+# ---------------------------------------------------------------------------
+
+
+class TestFetchAndMerge:
+    def _force_primary_disconnected(self):
+        """Set ais_stream module state so the gate allows the fallback."""
+        from services import ais_stream as ais
+        with ais._vessels_lock:
+            # Far in the past → connected = false; spawn_count > 0 → primary
+            # has at least tried so the gate engages.
+            ais._last_msg_at = time.time() - 3600
+            ais._proxy_spawn_count = 5
+            ais._vessels.clear()
+
+    def test_vessels_merged_with_source_tag(self, monkeypatch):
+        """Happy path: AISHub returns 2 ships, both land in ``_vessels``
+        with ``source: 'aishub'``."""
+        from services.fetchers import aishub_fallback
+        from services import ais_stream as ais
+
+        monkeypatch.setenv("AISHUB_USERNAME", "test-user")
+        self._force_primary_disconnected()
+
+        payload = json.dumps([
+            {"ERROR": False, "USERNAME": "test-user", "FORMAT": "1", "RECORDS": 2},
+            [
+                {
+                    "MMSI": 111111111,
+                    "LATITUDE": 40.0,
+                    "LONGITUDE": -73.0,
+                    "SOG": 12.0,
+                    "COG": 270,
+                    "HEADING": 275,
+                    "NAME": "SHIP A",
+                    "TYPE": 70,
+                },
+                {
+                    "MMSI": 222222222,
+                    "LATITUDE": 51.5,
+                    "LONGITUDE": -0.1,
+                    "SOG": 8.0,
+                    "COG": 90,
+                    "HEADING": 92,
+                    "NAME": "SHIP B",
+                    "TYPE": 60,
+                },
+            ],
+        ])
+
+        class FakeResp:
+            status_code = 200
+            text = payload
+
+        monkeypatch.setattr(
+            aishub_fallback, "fetch_with_curl", lambda *a, **kw: FakeResp()
+        )
+
+        count = aishub_fallback.fetch_aishub_vessels()
+
+        assert count == 2
+        with ais._vessels_lock:
+            v1 = ais._vessels.get(111111111)
+            v2 = ais._vessels.get(222222222)
+        assert v1 is not None
+        assert v1["source"] == "aishub"
+        assert v1["lat"] == 40.0
+        assert v1["name"] == "SHIP A"
+        assert v2 is not None
+        assert v2["source"] == "aishub"
+        assert v2["type"] == "passenger"  # AIS type 60 → passenger
+
+    def test_does_not_overwrite_fresh_live_data(self, monkeypatch):
+        """If the WebSocket pushed an update for an MMSI 0.5s ago and the
+        AISHub poll completes in that window, we should NOT clobber the
+        fresher live data."""
+        from services.fetchers import aishub_fallback
+        from services import ais_stream as ais
+
+        monkeypatch.setenv("AISHUB_USERNAME", "test-user")
+        self._force_primary_disconnected()
+
+        # Pre-seed _vessels with a "very fresh" live record.
+        fresh_ts = time.time()
+        with ais._vessels_lock:
+            ais._vessels[111111111] = {
+                "mmsi": 111111111,
+                "lat": 12.34,
+                "lng": 56.78,
+                "source": "aisstream",
+                "_updated": fresh_ts,
+            }
+
+        payload = json.dumps([
+            {"ERROR": False, "USERNAME": "test-user", "FORMAT": "1", "RECORDS": 1},
+            [
+                {
+                    "MMSI": 111111111,
+                    "LATITUDE": 99.0,  # bogus to make the test obvious
+                    "LONGITUDE": 99.0,
+                    "NAME": "STALE",
+                    "SOG": 0,
+                    "COG": 0,
+                    "TYPE": 0,
+                },
+            ],
+        ])
+
+        class FakeResp:
+            status_code = 200
+            text = payload
+
+        monkeypatch.setattr(
+            aishub_fallback, "fetch_with_curl", lambda *a, **kw: FakeResp()
+        )
+
+        # Note: 99.0/99.0 also exceeds the 91/181 sentinel guard and
+        # would be filtered. Pick a valid-but-bogus position instead.
+        payload = json.dumps([
+            {"ERROR": False, "USERNAME": "test-user", "FORMAT": "1", "RECORDS": 1},
+            [
+                {
+                    "MMSI": 111111111,
+                    "LATITUDE": 0.0,  # different from the live 12.34
+                    "LONGITUDE": 0.0,
+                    "NAME": "STALE",
+                    "SOG": 0,
+                    "COG": 0,
+                    "TYPE": 0,
+                },
+            ],
+        ])
+        monkeypatch.setattr(
+            aishub_fallback, "fetch_with_curl",
+            lambda *a, **kw: type("R", (), {"status_code": 200, "text": payload})(),
+        )
+
+        aishub_fallback.fetch_aishub_vessels()
+
+        with ais._vessels_lock:
+            v = ais._vessels.get(111111111)
+        # Live data wins — position should still be 12.34 / 56.78.
+        assert v["lat"] == 12.34
+        assert v["lng"] == 56.78
+        assert v["source"] == "aisstream"
+
+    def test_http_failure_returns_zero(self, monkeypatch):
+        from services.fetchers import aishub_fallback
+
+        monkeypatch.setenv("AISHUB_USERNAME", "test-user")
+        self._force_primary_disconnected()
+
+        class FailResp:
+            status_code = 503
+            text = ""
+
+        monkeypatch.setattr(
+            aishub_fallback, "fetch_with_curl", lambda *a, **kw: FailResp()
+        )
+
+        assert aishub_fallback.fetch_aishub_vessels() == 0

backend/tests/test_dm_relay_per_sender_cap.py

@@ -0,0 +1,270 @@
+"""Per-(sender, recipient) anti-spam cap on the DM relay.
+
+The user-stated rule: a single sender can have at most N UNACKED messages
+parked in a single recipient's mailbox at any one time (N=2 by default).
+Once the recipient pulls a message, the sender's quota for that pair
+frees up.
+
+Network rule, not local rule
+-----------------------------
+The cap is enforced TWICE:
+
+1. ``DMRelay.deposit(...)`` -- local check on the sender's own node.
+   Refuses to spool the (N+1)th message before it can be replicated.
+
+2. ``DMRelay.accept_replica(...)`` -- replication-acceptance check on
+   every receiving peer. Refuses to accept an inbound replica that
+   would put the local mailbox over the cap, even if the originating
+   peer claims it had cap room.
+
+The double enforcement matters because cap (1) is client-side -- a
+hostile relay could patch it out and continue to spool extras locally.
+Cap (2) means those extras can't propagate: every honest peer rejects
+them on the way in. A recipient who polls from honest peers therefore
+never sees more than N pending from any one sender, regardless of how
+many spam attempts the sender's own relay accepted.
+
+These tests pin both halves of the rule.
+"""
+
+from __future__ import annotations
+
+import time
+
+import pytest
+
+
+@pytest.fixture
+def relay():
+    """Fresh ``DMRelay`` per test."""
+    from services.mesh.mesh_dm_relay import DMRelay
+    r = DMRelay()
+    r._mailboxes.clear()
+    r._blocks.clear()
+    r._stats = {"messages_in_memory": 0}
+    return r
+
+
+def _deposit(
+    relay,
+    *,
+    sender: str = "alice",
+    recipient_token: str = "bob_mailbox_token_abc",
+    ciphertext: str = "ciphertext-blob",
+    msg_id: str = "",
+):
+    """Convenience wrapper using ``shared`` delivery class."""
+    return relay.deposit(
+        sender_id=sender,
+        raw_sender_id=sender,
+        recipient_id="bob",
+        ciphertext=ciphertext,
+        msg_id=msg_id,
+        delivery_class="shared",
+        recipient_token=recipient_token,
+    )
+
+
+# ---------------------------------------------------------------------------
+# Local cap on ``deposit``
+# ---------------------------------------------------------------------------
+
+
+class TestDepositCap:
+    def test_two_deposits_from_same_sender_succeed(self, relay):
+        r1 = _deposit(relay)
+        r2 = _deposit(relay)
+        assert r1["ok"] is True
+        assert r2["ok"] is True
+        assert r1["msg_id"] != r2["msg_id"]
+
+    def test_third_deposit_from_same_sender_rejected(self, relay):
+        _deposit(relay)
+        _deposit(relay)
+        r3 = _deposit(relay)
+        assert r3["ok"] is False
+        detail = r3["detail"].lower()
+        assert "unread" in detail or "read your messages" in detail
+
+    def test_different_senders_have_independent_quotas(self, relay):
+        for _ in range(2):
+            assert _deposit(relay, sender="alice")["ok"] is True
+        for _ in range(2):
+            assert _deposit(relay, sender="carol")["ok"] is True
+        assert _deposit(relay, sender="carol")["ok"] is False
+
+    def test_different_recipients_have_independent_quotas(self, relay):
+        for _ in range(2):
+            assert _deposit(relay, sender="alice", recipient_token="bob_token")["ok"] is True
+        for _ in range(2):
+            assert _deposit(relay, sender="alice", recipient_token="dave_token")["ok"] is True
+
+    def test_ack_frees_quota(self, relay):
+        r1 = _deposit(relay)
+        _deposit(relay)
+        assert _deposit(relay)["ok"] is False
+
+        mailbox_key = relay._hashed_mailbox_token("bob_mailbox_token_abc")
+        relay._mailboxes[mailbox_key] = [
+            m for m in relay._mailboxes[mailbox_key]
+            if m.msg_id != r1["msg_id"]
+        ]
+        relay._stats["messages_in_memory"] = sum(
+            len(v) for v in relay._mailboxes.values()
+        )
+
+        r3 = _deposit(relay)
+        assert r3["ok"] is True, f"expected quota free after ack, got: {r3}"
+
+    def test_cap_is_env_tunable(self, relay, monkeypatch):
+        import services.mesh.mesh_dm_relay as mdr
+        monkeypatch.setattr(
+            mdr.DMRelay,
+            "_per_sender_pending_limit",
+            lambda self: 1,
+        )
+
+        assert _deposit(relay)["ok"] is True
+        assert _deposit(relay)["ok"] is False
+
+
+# ---------------------------------------------------------------------------
+# Replication-acceptance cap (the half that makes this a network rule)
+# ---------------------------------------------------------------------------
+
+
+class TestAcceptReplicaCap:
+    def _envelope(self, *, msg_id: str, sender_block_ref: str, mailbox_key: str):
+        return {
+            "msg_id": msg_id,
+            "mailbox_key": mailbox_key,
+            "sender_block_ref": sender_block_ref,
+            "sender_id": "alice",
+            "sender_seal": "",
+            "ciphertext": f"ciphertext-{msg_id}",
+            "timestamp": time.time(),
+            "delivery_class": "shared",
+            "relay_salt": "",
+            "payload_format": "dm1",
+            "session_welcome": "",
+        }
+
+    def test_replica_accepted_under_cap(self, relay):
+        env = self._envelope(
+            msg_id="dm_replica_1",
+            sender_block_ref="alice_block_ref",
+            mailbox_key="mailbox_xyz",
+        )
+        result = relay.accept_replica(envelope=env)
+        assert result["ok"] is True
+
+    def test_replica_idempotent_on_duplicate_msg_id(self, relay):
+        mailbox_key = "mailbox_xyz"
+        env = self._envelope(
+            msg_id="dm_dup_1",
+            sender_block_ref="alice_block_ref",
+            mailbox_key=mailbox_key,
+        )
+        r1 = relay.accept_replica(envelope=env)
+        r2 = relay.accept_replica(envelope=env)
+        assert r1["ok"] is True
+        assert r2["ok"] is True
+        assert r2.get("duplicate") is True
+        assert len(relay._mailboxes[mailbox_key]) == 1
+
+    def test_replica_rejected_when_local_count_already_at_cap(self, relay):
+        mailbox_key = "mailbox_xyz"
+        for i in (1, 2):
+            relay.accept_replica(envelope=self._envelope(
+                msg_id=f"dm_seeded_{i}",
+                sender_block_ref="alice_block_ref",
+                mailbox_key=mailbox_key,
+            ))
+
+        result = relay.accept_replica(envelope=self._envelope(
+            msg_id="dm_overcap_3",
+            sender_block_ref="alice_block_ref",
+            mailbox_key=mailbox_key,
+        ))
+        assert result["ok"] is False
+        assert result.get("cap_violation") is True
+        assert result.get("pending") == 2
+        assert result.get("limit") == 2
+        assert len(relay._mailboxes[mailbox_key]) == 2
+
+    def test_replica_from_different_sender_passes_when_one_is_at_cap(self, relay):
+        mailbox_key = "mailbox_xyz"
+        for i in (1, 2):
+            relay.accept_replica(envelope=self._envelope(
+                msg_id=f"dm_alice_{i}",
+                sender_block_ref="alice_block_ref",
+                mailbox_key=mailbox_key,
+            ))
+        assert relay.accept_replica(envelope=self._envelope(
+            msg_id="dm_alice_3",
+            sender_block_ref="alice_block_ref",
+            mailbox_key=mailbox_key,
+        ))["ok"] is False
+        assert relay.accept_replica(envelope=self._envelope(
+            msg_id="dm_carol_1",
+            sender_block_ref="carol_block_ref",
+            mailbox_key=mailbox_key,
+        ))["ok"] is True
+
+    def test_replica_rejects_malformed_envelopes(self, relay):
+        for bad in (
+            {},
+            {"msg_id": "x"},
+            {"msg_id": "x", "mailbox_key": "y"},
+            "not an object at all",
+        ):
+            result = relay.accept_replica(envelope=bad)
+            assert result["ok"] is False
+
+
+# ---------------------------------------------------------------------------
+# ``envelope_for_replication`` -- helper for the outbound replication path
+# ---------------------------------------------------------------------------
+
+
+class TestEnvelopeForReplication:
+    def test_returns_envelope_for_stored_message(self, relay):
+        r = _deposit(relay, ciphertext="hello-ciphertext")
+        msg_id = r["msg_id"]
+        mailbox_key = relay._hashed_mailbox_token("bob_mailbox_token_abc")
+
+        env = relay.envelope_for_replication(mailbox_key=mailbox_key, msg_id=msg_id)
+        assert env is not None
+        assert env["msg_id"] == msg_id
+        assert env["mailbox_key"] == mailbox_key
+        assert env["ciphertext"] == "hello-ciphertext"
+        assert env["delivery_class"] == "shared"
+        for k in ("msg_id", "mailbox_key", "sender_block_ref", "ciphertext"):
+            assert env.get(k), f"envelope missing required field {k!r}"
+
+    def test_returns_none_for_unknown_message(self, relay):
+        env = relay.envelope_for_replication(
+            mailbox_key="never_existed", msg_id="never_existed",
+        )
+        assert env is None
+
+    def test_envelope_round_trips_through_accept_replica(self, relay):
+        from services.mesh.mesh_dm_relay import DMRelay
+        receiver_relay = DMRelay()
+        receiver_relay._mailboxes.clear()
+        receiver_relay._stats = {"messages_in_memory": 0}
+
+        r = _deposit(relay)
+        msg_id = r["msg_id"]
+        mailbox_key = relay._hashed_mailbox_token("bob_mailbox_token_abc")
+        env = relay.envelope_for_replication(
+            mailbox_key=mailbox_key, msg_id=msg_id,
+        )
+        assert env is not None
+
+        result = receiver_relay.accept_replica(envelope=env)
+        assert result["ok"] is True
+        stored = receiver_relay._mailboxes.get(mailbox_key, [])
+        assert len(stored) == 1
+        assert stored[0].msg_id == msg_id
+        assert stored[0].ciphertext == "ciphertext-blob"

backend/tests/test_dm_replicate_envelope_endpoint.py

@@ -0,0 +1,150 @@
+"""POST /api/mesh/dm/replicate-envelope — receiving side of cross-node DM
+mailbox replication.
+
+This is the endpoint that peer relays call when they want to hand off an
+encrypted DM envelope to us (so the recipient can log into our node and
+find their messages). It re-enforces the per-(sender, recipient) anti-spam
+cap so hostile sender relays can't widen the cap by skipping the local
+check on their own deposit path.
+
+The endpoint:
+
+  * authenticates the caller via the existing per-peer HMAC pattern
+    (same one /api/mesh/infonet/peer-push and /api/mesh/gate/peer-push
+    use, introduced in #256 — ``X-Peer-Url`` + ``X-Peer-HMAC`` headers
+    keyed off ``resolve_peer_key_for_url``)
+  * rejects bodies > 64 KB (DM envelope size is bounded by
+    ``MESH_DM_MAX_MSG_BYTES`` — 64KB ceiling has generous headroom)
+  * rejects requests without a valid peer HMAC with 403
+  * passes the envelope to ``DMRelay.accept_replica`` which enforces
+    the cap
+
+This file pins the endpoint contract. The cap enforcement itself is
+tested in ``test_dm_relay_per_sender_cap.py`` against the relay's
+``accept_replica`` method directly.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import hashlib
+import hmac
+import json
+
+import pytest
+from httpx import ASGITransport, AsyncClient
+
+
+@pytest.fixture
+def remote_client():
+    """ASGI client with peer IP 1.2.3.4 — never on the local-operator
+    allowlist. Used to prove the endpoint isn't accidentally reachable
+    by random remote callers without peer HMAC."""
+    from main import app
+
+    class _RemoteClient:
+        def __init__(self):
+            self._loop = asyncio.new_event_loop()
+            self._transport = ASGITransport(app=app, client=("1.2.3.4", 12345))
+            self._base = "http://1.2.3.4:8000"
+
+        def post(self, url, **kw):
+            async def go():
+                async with AsyncClient(transport=self._transport, base_url=self._base) as ac:
+                    return await ac.post(url, **kw)
+            return self._loop.run_until_complete(go())
+
+        def close(self):
+            self._loop.close()
+
+    c = _RemoteClient()
+    yield c
+    c.close()
+
+
+class TestReplicateEndpointAuth:
+    def test_rejects_request_without_peer_hmac(self, remote_client):
+        """A peer push that does NOT carry X-Peer-Url + X-Peer-HMAC
+        must be rejected with 403 before the envelope is ever passed
+        to the relay. Same gate the existing infonet/gate peer-push
+        endpoints enforce."""
+        payload = {
+            "envelope": {
+                "msg_id": "dm_unauth_1",
+                "mailbox_key": "mb",
+                "sender_block_ref": "sender",
+                "ciphertext": "x",
+            },
+        }
+        r = remote_client.post(
+            "/api/mesh/dm/replicate-envelope",
+            json=payload,
+        )
+        assert r.status_code == 403
+        assert "peer HMAC" in r.text or "peer hmac" in r.text.lower()
+
+    def test_rejects_wrong_peer_hmac(self, remote_client, monkeypatch):
+        """A request with a peer HMAC header keyed off the WRONG secret
+        is rejected. Confirms the HMAC is actually verified — a tampered
+        body or a key-substitution attack doesn't sneak through."""
+        # Plant a known peer secret. The request will sign with a
+        # DIFFERENT key, so verification must fail.
+        from services.config import get_settings
+        monkeypatch.setenv("MESH_PEER_PUSH_SECRET", "real-secret-32-chars-min-padding-padding")
+        get_settings.cache_clear()
+
+        body = json.dumps({
+            "envelope": {
+                "msg_id": "dm_wronghmac",
+                "mailbox_key": "mb",
+                "sender_block_ref": "sender",
+                "ciphertext": "x",
+            },
+        }).encode("utf-8")
+        wrong_hmac = hmac.new(b"wrong-key", body, hashlib.sha256).hexdigest()
+        r = remote_client.post(
+            "/api/mesh/dm/replicate-envelope",
+            content=body,
+            headers={
+                "Content-Type": "application/json",
+                "X-Peer-Url": "http://example-peer.onion:8000",
+                "X-Peer-HMAC": wrong_hmac,
+            },
+        )
+        assert r.status_code == 403
+
+    def test_rejects_oversize_body(self, remote_client):
+        """64 KB ceiling — anything bigger doesn't even get parsed.
+        Defends against memory amplification via giant ciphertexts."""
+        # 100 KB body is well over the 64 KB cap.
+        big = b"{" + b"x" * 100_000 + b"}"
+        r = remote_client.post(
+            "/api/mesh/dm/replicate-envelope",
+            content=big,
+            headers={
+                "Content-Type": "application/json",
+                "Content-Length": str(len(big)),
+            },
+        )
+        assert r.status_code in (400, 413), (
+            f"oversize body should be rejected with 400/413, got {r.status_code}"
+        )
+
+
+class TestReplicateEndpointRegistered:
+    def test_route_present_in_app(self):
+        """Static check that the route is actually wired into the app.
+        Catches a future refactor that drops the router include or
+        deletes the endpoint by accident."""
+        from main import app
+
+        paths_methods = set()
+        for route in app.routes:
+            path = getattr(route, "path", None)
+            methods = getattr(route, "methods", set()) or set()
+            for m in methods:
+                paths_methods.add((m, path))
+
+        assert ("POST", "/api/mesh/dm/replicate-envelope") in paths_methods, (
+            "POST /api/mesh/dm/replicate-envelope is not registered on the app"
+        )

backend/tests/test_flight_observations.py

@@ -0,0 +1,258 @@
+"""Cumulative fuel/CO2 tracking via per-aircraft observation timestamps.
+
+Background
+----------
+Users want the running total of fuel burned per aircraft — not just the
+rate. We track first-seen-at per icao24 and multiply elapsed observation
+time by the model-based rate. This module's job is exclusively the
+timestamp bookkeeping; multiplication happens in the flights/military
+fetchers.
+
+These tests pin:
+
+  * First sighting returns 0 (no airtime yet).
+  * Repeated sightings within ``REOPEN_GAP_S`` accumulate elapsed time.
+  * Gap longer than ``REOPEN_GAP_S`` resets the session (plane landed
+    and took off again — different flight).
+  * ``MAX_SESSION_SECONDS`` clamp protects against clock skew bugs.
+  * ``prune()`` drops stale entries.
+  * ``get_session_seconds`` reads without bumping last_seen.
+  * Empty / None icao input is a defensive no-op.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+
+@pytest.fixture(autouse=True)
+def _reset_observations():
+    from services.fetchers import flight_observations as obs
+    obs._reset_for_tests()
+    yield
+    obs._reset_for_tests()
+
+
+class TestRecordObservation:
+    def test_first_sighting_returns_zero(self):
+        from services.fetchers.flight_observations import record_observation
+        assert record_observation("a12345", now=1000.0) == 0
+
+    def test_repeated_sightings_accumulate(self):
+        """ADS-B refreshes every ~minute in practice, so each observation
+        is within ``REOPEN_GAP_S`` (15 min) of the last and we keep
+        accumulating. Walking the timestamps in 5-minute steps so we
+        stay inside the reopen window the whole way."""
+        from services.fetchers.flight_observations import record_observation
+        record_observation("a12345", now=1000.0)
+        # 1 minute later (within REOPEN_GAP_S)
+        assert record_observation("a12345", now=1060.0) == 60
+        # Step through 5-minute spaced refreshes — first_seen_at stays
+        # at 1000.0 the whole time, and we approach a 1-hour airtime.
+        assert record_observation("a12345", now=1360.0) == 360
+        assert record_observation("a12345", now=1660.0) == 660
+        assert record_observation("a12345", now=1960.0) == 960
+        assert record_observation("a12345", now=2260.0) == 1260
+        assert record_observation("a12345", now=2560.0) == 1560
+        assert record_observation("a12345", now=2860.0) == 1860
+        assert record_observation("a12345", now=3160.0) == 2160
+        assert record_observation("a12345", now=3460.0) == 2460
+        assert record_observation("a12345", now=3760.0) == 2760
+        assert record_observation("a12345", now=4060.0) == 3060
+        assert record_observation("a12345", now=4360.0) == 3360
+        # 1 hour after first sighting — still inside the 15-min reopen
+        # window from the prior 4360 observation.
+        assert record_observation("a12345", now=4600.0) == 3600
+
+    def test_gap_longer_than_reopen_resets_session(self):
+        """If a hex hasn't been seen in ``REOPEN_GAP_S`` (15 min default),
+        the next sighting is treated as a new flight — first_seen_at resets."""
+        from services.fetchers.flight_observations import record_observation
+        record_observation("a12345", now=1000.0)
+        record_observation("a12345", now=1500.0)  # 500s later — within gap
+        # Now 20 minutes of silence (1200s > 900s threshold) → session reset.
+        assert record_observation("a12345", now=2700.0) == 0
+        # And the next quick sighting starts accumulating from 2700 again.
+        assert record_observation("a12345", now=2760.0) == 60
+
+    def test_session_clamp(self):
+        """Clock skew protection: when a hex has been continuously
+        observed for longer than ``MAX_SESSION_SECONDS``, clamp.
+
+        Synthesizes the state directly because driving 86,400+ seconds of
+        observations through the public API in a test would take 1000+
+        REOPEN_GAP_S-respecting steps.
+        """
+        from services.fetchers import flight_observations as obs
+        from services.fetchers.flight_observations import _observations, _lock
+
+        # last_seen_at very recent so REOPEN_GAP_S branch does NOT fire,
+        # but first_seen_at way in the past so the elapsed math overflows
+        # MAX_SESSION_SECONDS. Clamp must kick in.
+        big_now = float(obs.MAX_SESSION_SECONDS + 1_000_000)
+        with _lock:
+            _observations["a12345"] = {
+                "first_seen_at": 0.0,
+                "last_seen_at": big_now - 60,  # 60s ago — well inside gap window
+            }
+        elapsed = obs.record_observation("a12345", now=big_now)
+        assert elapsed == obs.MAX_SESSION_SECONDS, (
+            f"elapsed must be clamped to MAX_SESSION_SECONDS; got {elapsed}"
+        )
+
+    def test_empty_input_returns_zero(self):
+        from services.fetchers.flight_observations import record_observation
+        assert record_observation("") == 0
+        assert record_observation(None) == 0  # type: ignore[arg-type]
+        assert record_observation("   ") == 0
+
+    def test_case_insensitive_key(self):
+        """ICAO24 hex codes are case-insensitive — adsb.lol lowercases
+        them, OpenSky may not. Normalize so both refer to the same airframe."""
+        from services.fetchers.flight_observations import record_observation
+        record_observation("A12345", now=1000.0)
+        # Different case must hit the same entry.
+        assert record_observation("a12345", now=1060.0) == 60
+
+
+class TestGetSessionSeconds:
+    def test_read_only_does_not_bump(self):
+        from services.fetchers.flight_observations import (
+            record_observation,
+            get_session_seconds,
+        )
+        record_observation("a12345", now=1000.0)
+        record_observation("a12345", now=1060.0)  # bumps last_seen
+
+        # Now read at t=2000. Without bumping, gap=2000-1060=940 > 900,
+        # so a recording call would reset. But the read should NOT reset.
+        seconds_at_2000 = get_session_seconds("a12345", now=2000.0)
+        assert seconds_at_2000 == 1000, (
+            f"read should return 2000-1000=1000s; got {seconds_at_2000}"
+        )
+        # Verify the next recording at t=2001 still resets (gap > 900s
+        # from the read above — proves the read didn't bump last_seen).
+        from services.fetchers.flight_observations import record_observation as rec
+        assert rec("a12345", now=2001.0) == 0  # session reset
+
+    def test_unknown_hex_returns_zero(self):
+        from services.fetchers.flight_observations import get_session_seconds
+        assert get_session_seconds("nonexistent") == 0
+
+
+class TestPrune:
+    def test_drops_stale_entries(self):
+        from services.fetchers import flight_observations as obs
+
+        obs.record_observation("active", now=10_000.0)
+        obs.record_observation("stale", now=1.0)
+
+        dropped = obs.prune(now=10_000.0)
+
+        assert dropped == 1
+        # Active entry survives:
+        assert obs.get_session_seconds("active", now=10_001.0) == 1
+        # Stale entry was dropped — next obs starts fresh:
+        assert obs.record_observation("stale", now=10_002.0) == 0
+
+    def test_no_op_when_nothing_stale(self):
+        from services.fetchers import flight_observations as obs
+        obs.record_observation("hex1", now=1000.0)
+        obs.record_observation("hex2", now=1000.0)
+
+        dropped = obs.prune(now=1500.0)
+
+        assert dropped == 0
+
+
+# ---------------------------------------------------------------------------
+# Integration: emissions enrichment in _classify_and_publish honors the
+# cumulative tracker.
+# ---------------------------------------------------------------------------
+
+
+class TestEmissionsCumulativeIntegration:
+    def _reset_store(self):
+        from services.fetchers._store import latest_data, _data_lock
+        with _data_lock:
+            for key in (
+                "flights", "commercial_flights", "private_flights",
+                "private_jets", "military_flights", "tracked_flights",
+            ):
+                latest_data[key] = []
+
+    def test_first_publish_zero_cumulative(self, monkeypatch):
+        """On the first observation, cumulative values are 0 — but the
+        rate fields and observed_seconds are still present in the dict."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish([
+            {
+                "hex": "test001",
+                "flight": "JBU711",
+                "r": "N1",
+                "t": "C172",  # Cessna 172, 9 GPH
+                "lat": 40.0,
+                "lon": -100.0,
+                "alt_baro": 3000,
+                "gs": 100,
+            }
+        ])
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert len(published) == 1
+        emi = published[0].get("emissions")
+        assert emi is not None
+        assert emi["fuel_gph"] == 9
+        assert emi["observed_seconds"] == 0
+        assert emi["fuel_gallons_burned"] == 0.0
+        assert emi["co2_kg_emitted"] == 0.0
+
+    def test_second_publish_accumulates(self, monkeypatch):
+        """Publishing the same hex a second time picks up real elapsed time
+        and produces non-zero cumulative values."""
+        import time as _time_real
+        from services.fetchers import flights as flights_module
+        from services.fetchers import flight_observations as obs
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        # Manually seed an observation 1 hour in the past so the next
+        # publish picks up ~3600s elapsed.
+        with obs._lock:
+            obs._observations["test002"] = {
+                "first_seen_at": _time_real.time() - 3600,
+                "last_seen_at": _time_real.time() - 60,
+            }
+
+        flights_module._classify_and_publish([
+            {
+                "hex": "test002",
+                "flight": "JBU711",
+                "r": "N1",
+                "t": "C172",  # 9 GPH
+                "lat": 40.0,
+                "lon": -100.0,
+                "alt_baro": 3000,
+                "gs": 100,
+            }
+        ])
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert len(published) == 1
+        emi = published[0].get("emissions")
+        # Roughly 1 hour observed → 9 gal burned.
+        assert 3500 <= emi["observed_seconds"] <= 3700
+        assert 8.7 <= emi["fuel_gallons_burned"] <= 9.3
+        # CO2 = 9 gph * 9.57 kg/gal = 86.1 kg/hr.
+        assert 84 <= emi["co2_kg_emitted"] <= 88

backend/tests/test_flight_source_attribution.py

@@ -0,0 +1,354 @@
+"""Per-flight source attribution.
+
+Background
+----------
+Pre-fix, adsb.lol records (the primary source for most flights) carried
+no source marker. OpenSky records got ``is_opensky: True`` and
+supplementals got ``supplemental_source``, so any UI that wanted to show
+which provider a flight came from saw OpenSky/airplanes.live records as
+explicitly tagged and adsb.lol records as "unlabeled" — making it look
+like adsb.lol wasn't even being used.
+
+This caused user confusion ("only military planes have adsb.lol
+telemetry") that was diagnostic noise, not a real bug. The actual fix:
+stamp ``source`` at every fetch site so the downstream consumer can
+attribute the provider with no guesswork.
+
+These tests pin:
+
+  * adsb.lol regional records get ``source: "adsb.lol"`` at fetch time
+    (synthesized via the published flight dict).
+  * OpenSky records get ``source: "OpenSky"`` (alongside the existing
+    ``is_opensky: True`` for backwards compat).
+  * Supplementals (airplanes.live, adsb.fi) flow through with their
+    ``supplemental_source`` honored.
+  * The military fetcher tags ``source`` on military_flights and uavs.
+  * The published flight dict carries ``source`` so downstream code
+    can render attribution.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+
+# ---------------------------------------------------------------------------
+# _classify_and_publish — source field flows into published flight dict
+# ---------------------------------------------------------------------------
+
+
+class TestClassifyAndPublishSource:
+    def _reset_store(self):
+        """Clear store before each test so we get deterministic state."""
+        from services.fetchers._store import latest_data, _data_lock
+        with _data_lock:
+            for key in (
+                "flights", "commercial_flights", "private_flights",
+                "private_jets", "military_flights", "tracked_flights",
+            ):
+                latest_data[key] = []
+        return latest_data
+
+    def test_adsb_lol_record_tagged_in_published_flight(self, monkeypatch):
+        """A raw adsb.lol record (carrying ``source: 'adsb.lol'`` from the
+        fetch site) flows through ``_classify_and_publish`` and the
+        published flight dict carries the same ``source`` field."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+
+        # Patch route + type lookups so they don't try to hit the network.
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish(
+            [
+                {
+                    "hex": "ad7701",
+                    "flight": "JBU711",
+                    "r": "N967JT",
+                    "t": "A321",
+                    "lat": 40.0,
+                    "lon": -100.0,
+                    "alt_baro": 36000,
+                    "gs": 401.6,
+                    "nac_p": 9,
+                    "source": "adsb.lol",  # stamped at fetch site
+                }
+            ]
+        )
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert len(published) == 1
+        assert published[0]["source"] == "adsb.lol"
+        # nac_p still flows through too — sanity check that adding source
+        # didn't break the existing GPS jamming signal.
+        assert published[0]["nac_p"] == 9
+
+    def test_opensky_record_tagged_in_published_flight(self, monkeypatch):
+        """OpenSky-sourced records carry ``source: 'OpenSky'`` (plus the
+        existing ``is_opensky: True`` for back-compat)."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish(
+            [
+                {
+                    "hex": "a12345",
+                    "flight": "UAL100",
+                    "r": "N100UA",
+                    "t": "Unknown",
+                    "lat": 41.0,
+                    "lon": -87.0,
+                    "alt_baro": 35000,
+                    "gs": 450,
+                    # No nac_p — OpenSky doesn't carry it.
+                    "is_opensky": True,
+                    "source": "OpenSky",
+                }
+            ]
+        )
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert len(published) == 1
+        assert published[0]["source"] == "OpenSky"
+
+    def test_supplemental_source_propagates(self, monkeypatch):
+        """Supplemental records (airplanes.live, adsb.fi) have their
+        legacy ``supplemental_source`` field promoted to the unified
+        ``source`` field in the published dict — so consumers don't have
+        to inspect two different keys."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish(
+            [
+                {
+                    "hex": "b22222",
+                    "flight": "DAL200",
+                    "r": "N200DL",
+                    "t": "B738",
+                    "lat": 42.0,
+                    "lon": -90.0,
+                    "alt_baro": 32000,
+                    "gs": 420,
+                    "supplemental_source": "airplanes.live",
+                    # No explicit "source" — should fall through to
+                    # supplemental_source.
+                }
+            ]
+        )
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert len(published) == 1
+        assert published[0]["source"] == "airplanes.live"
+
+    def test_explicit_source_wins_over_supplemental_source(self, monkeypatch):
+        """If both fields are present, explicit ``source`` wins (it's the
+        newer canonical tag)."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish(
+            [
+                {
+                    "hex": "c33333",
+                    "flight": "AAL300",
+                    "r": "N300AA",
+                    "t": "A321",
+                    "lat": 33.0,
+                    "lon": -97.0,
+                    "alt_baro": 34000,
+                    "gs": 430,
+                    "source": "adsb.lol",
+                    "supplemental_source": "adsb.fi",
+                }
+            ]
+        )
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert published[0]["source"] == "adsb.lol"
+
+    def test_untagged_record_defaults_to_adsb_lol(self, monkeypatch):
+        """A record with neither ``source`` nor ``supplemental_source``
+        (e.g. synthesized by a test, or a fetcher that hasn't been
+        migrated yet) defaults to ``"adsb.lol"`` since that's been the
+        primary source historically. Defensive default — better than
+        empty string."""
+        from services.fetchers import flights as flights_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        self._reset_store()
+        monkeypatch.setattr(flights_module, "lookup_route", lambda _: None)
+        monkeypatch.setattr(flights_module, "lookup_aircraft_type", lambda _: "")
+
+        flights_module._classify_and_publish(
+            [
+                {
+                    "hex": "d44444",
+                    "flight": "SWA400",
+                    "r": "N400SW",
+                    "t": "B737",
+                    "lat": 32.0,
+                    "lon": -110.0,
+                    "alt_baro": 30000,
+                    "gs": 410,
+                }
+            ]
+        )
+
+        with _data_lock:
+            published = list(latest_data.get("flights", []))
+        assert published[0]["source"] == "adsb.lol"
+
+
+# ---------------------------------------------------------------------------
+# adsb.lol regional fetcher tags at fetch time
+# ---------------------------------------------------------------------------
+
+
+class TestAdsbLolRegionalTagging:
+    def test_fetch_region_stamps_source_on_each_aircraft(self, monkeypatch):
+        """The wrapper around the adsb.lol regional endpoint stamps
+        ``source: 'adsb.lol'`` on every record before returning, so the
+        downstream merge step sees attribution survive even when the
+        record gets reshuffled (e.g. dedupe-by-hex during OpenSky merge)."""
+        from services.fetchers import flights as flights_module
+
+        # Fake response — 3 aircraft, none have a source field originally.
+        class FakeResp:
+            status_code = 200
+
+            def json(self):
+                return {
+                    "ac": [
+                        {"hex": "a1", "lat": 40.0, "lon": -100.0, "nac_p": 8},
+                        {"hex": "a2", "lat": 40.1, "lon": -100.1, "nac_p": 9},
+                        {"hex": "a3", "lat": 40.2, "lon": -100.2, "nac_p": 10},
+                    ]
+                }
+
+        monkeypatch.setattr(
+            flights_module, "fetch_with_curl", lambda *a, **kw: FakeResp()
+        )
+
+        results = flights_module._fetch_adsb_lol_regions()
+
+        assert len(results) >= 3
+        # Every aircraft we got back must be tagged.
+        sources = {a.get("source") for a in results}
+        assert sources == {"adsb.lol"}, (
+            f"adsb.lol regional fetcher must stamp source on every record; "
+            f"got: {sources}"
+        )
+
+    def test_fetch_region_failure_returns_empty_without_crashing(self, monkeypatch):
+        """If adsb.lol returns non-200, the fetcher returns [] gracefully —
+        downstream code already handles this. Sanity check that the source
+        tagging doesn't introduce a new failure mode."""
+        from services.fetchers import flights as flights_module
+
+        class FakeResp:
+            status_code = 500
+            def json(self): return {}
+
+        monkeypatch.setattr(
+            flights_module, "fetch_with_curl", lambda *a, **kw: FakeResp()
+        )
+
+        results = flights_module._fetch_adsb_lol_regions()
+
+        assert results == []
+
+
+# ---------------------------------------------------------------------------
+# Military fetcher tags source on output dicts
+# ---------------------------------------------------------------------------
+
+
+class TestMilitarySourceTagging:
+    def test_military_output_carries_source_field(self, monkeypatch):
+        """Each entry in ``military_flights`` should carry a ``source``
+        field. Pre-fix the only military attribution was inferring from
+        which endpoint we hit; now it's explicit."""
+        from services.fetchers import military as mil_module
+        from services.fetchers._store import latest_data, _data_lock
+
+        # Reset relevant store state.
+        with _data_lock:
+            latest_data["military_flights"] = []
+            latest_data["uavs"] = []
+            latest_data["tracked_flights"] = []
+
+        # Stub _store.is_any_active so the fetch doesn't early-return.
+        # The military module imports the function inline at call time,
+        # so we have to patch it on the _store module itself rather than
+        # on the military module.
+        from services.fetchers import _store as store_module
+        monkeypatch.setattr(store_module, "is_any_active", lambda *_: True)
+
+        # Stub fetch_with_curl to return one synthetic military aircraft
+        # from adsb.lol, none from airplanes.live.
+        class _RespMil:
+            status_code = 200
+            def json(self):
+                return {
+                    "ac": [
+                        {
+                            "hex": "ae6c1d",
+                            "flight": "CRUSH52",
+                            "r": "170281",
+                            "t": "C30J",
+                            "lat": 47.594,
+                            "lon": -124.879,
+                            "alt_baro": 9025,
+                            "gs": 162.8,
+                            "track": 334.5,
+                            "nac_p": 10,
+                        }
+                    ]
+                }
+
+        class _RespEmpty:
+            status_code = 200
+            def json(self):
+                return {"ac": []}
+
+        def _fake_fetch(url, *a, **kw):
+            if "adsb.lol" in url:
+                return _RespMil()
+            return _RespEmpty()
+
+        monkeypatch.setattr(mil_module, "fetch_with_curl", _fake_fetch)
+        # Stubs for downstream enrichments that try to hit external state.
+        monkeypatch.setattr(mil_module, "enrich_with_plane_alert", lambda mf: None)
+        monkeypatch.setattr(mil_module, "_enrich_country", lambda hex_, flag: ("US", "USAF"))
+        monkeypatch.setattr(mil_module, "_classify_military_type", lambda t: "transport")
+        monkeypatch.setattr(mil_module, "_classify_uav", lambda m, c: (False, "", ""))
+        monkeypatch.setattr(mil_module, "get_emissions_info", lambda model: None)
+        monkeypatch.setattr(mil_module, "_mark_fresh", lambda *keys: None)
+
+        mil_module.fetch_military_flights()
+
+        with _data_lock:
+            mil_published = list(latest_data.get("military_flights", []))
+
+        assert len(mil_published) == 1
+        assert mil_published[0]["source"] == "adsb.lol"

backend/tests/test_gps_jamming_detector.py

@@ -0,0 +1,333 @@
+"""GPS jamming detection — nac_p=0 counted, lowered thresholds.
+
+Background
+----------
+Pre-fix, the detector had three stacked filters that together meant the
+``gps_jamming`` layer almost never lit up:
+
+  1. ``nac_p == 0`` aircraft were dropped on the theory that "0 = old
+     transponder." But modern Mode-S Enhanced Surveillance transponders
+     also fall back to ``nac_p == 0`` when they lose GPS lock entirely —
+     which is *exactly* the jamming signature we want to catch.
+  2. ``GPS_JAMMING_MIN_AIRCRAFT = 5`` per 1°x1° cell.
+  3. ``GPS_JAMMING_MIN_RATIO = 0.30`` adjusted ratio.
+
+Combined with the existing ``-1`` noise cushion (``adjusted = degraded - 1``)
+the bar to clear required dense, busy airspace — but jamming hotspots
+(eastern Med, eastern Ukraine, Iran/Iraq) tend to have sparser traffic
+precisely because pilots avoid them.
+
+These tests pin the new behavior:
+
+  * ``nac_p == 0`` is now counted as degraded.
+  * ``nac_p == None`` (no field — typical for OpenSky records) is still
+    skipped — absence isn't evidence.
+  * Thresholds lowered to 3 aircraft / 0.20 ratio.
+  * Public function signature accepts overrides so callers / future
+    operators can re-tune without code edits.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+
+# ---------------------------------------------------------------------------
+# nac_p == 0 inclusion (the headline fix)
+# ---------------------------------------------------------------------------
+
+
+class TestNacpZeroCounted:
+    def test_cell_dominated_by_nacp_zero_now_fires(self):
+        """Three aircraft all reporting nac_p=0 in one cell, plus two
+        with valid GPS. Pre-fix the three nac_p=0 records were skipped
+        entirely (cell would have total=2, degraded=0, no zone). Post-fix
+        they count as degraded — this IS the jamming signature."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        # All in 1°x1° cell at int(lat)=40, int(lng)=-100
+        feed = [
+            {"hex": "a1", "lat": 40.1, "lng": -100.1, "nac_p": 0},
+            {"hex": "a2", "lat": 40.5, "lng": -100.5, "nac_p": 0},
+            {"hex": "a3", "lat": 40.9, "lng": -100.9, "nac_p": 0},
+            {"hex": "b1", "lat": 40.2, "lng": -100.3, "nac_p": 9},
+            {"hex": "b2", "lat": 40.7, "lng": -100.7, "nac_p": 11},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # total=5, degraded=3, adjusted=2, ratio=0.40 > 0.20 → zone fires.
+        assert len(zones) == 1
+        assert zones[0]["degraded"] == 3
+        assert zones[0]["total"] == 5
+        assert zones[0]["ratio"] == 0.40
+        # Grid-cell center coords.
+        assert zones[0]["lat"] == 40.5
+        assert zones[0]["lng"] == -99.5
+
+    def test_nacp_zero_alone_clears_min_aircraft(self):
+        """A cell with exactly 3 aircraft all reporting nac_p=0 must
+        fire under the new MIN_AIRCRAFT=3 + MIN_RATIO=0.20 regime."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 50.1, "lng": 30.1, "nac_p": 0},
+            {"hex": "a2", "lat": 50.5, "lng": 30.5, "nac_p": 0},
+            {"hex": "a3", "lat": 50.9, "lng": 30.9, "nac_p": 0},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # total=3, degraded=3, adjusted=2, ratio=0.667 > 0.20 → fires.
+        # severity is "medium" because 0.5 ≤ ratio < 0.75.
+        assert len(zones) == 1
+        assert zones[0]["severity"] == "medium"
+
+
+# ---------------------------------------------------------------------------
+# nac_p == None is still skipped (preserve OpenSky behavior)
+# ---------------------------------------------------------------------------
+
+
+class TestNoneStillSkipped:
+    def test_none_records_dont_add_to_grid(self):
+        """OpenSky's /states/all doesn't include nac_p, so its records
+        arrive with the field absent (``rf.get("nac_p") is None``). These
+        records must NOT count toward total — absence-of-data isn't
+        evidence of either jamming OR working GPS."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        # 3 jammed + 4 OpenSky-style (no nac_p). Pre-fix and post-fix
+        # behavior should be identical here: None always skipped.
+        feed = [
+            {"hex": "a1", "lat": 40.1, "lng": -100.1, "nac_p": 0},
+            {"hex": "a2", "lat": 40.2, "lng": -100.2, "nac_p": 0},
+            {"hex": "a3", "lat": 40.3, "lng": -100.3, "nac_p": 0},
+            # OpenSky-style: no nac_p at all
+            {"hex": "o1", "lat": 40.4, "lng": -100.4},
+            {"hex": "o2", "lat": 40.5, "lng": -100.5},
+            {"hex": "o3", "lat": 40.6, "lng": -100.6},
+            {"hex": "o4", "lat": 40.7, "lng": -100.7},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # Only the 3 nac_p=0 records hit the grid. total=3, not 7.
+        assert len(zones) == 1
+        assert zones[0]["total"] == 3
+        assert zones[0]["degraded"] == 3
+
+    def test_explicit_none_skipped(self):
+        """Same behavior when ``nac_p`` is present but set to None
+        (defensive — adsb.lol shouldn't do this, but downstream
+        normalizers might)."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 0.1, "lng": 0.1, "nac_p": None},
+            {"hex": "a2", "lat": 0.2, "lng": 0.2, "nac_p": None},
+            {"hex": "a3", "lat": 0.3, "lng": 0.3, "nac_p": None},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # No records counted → no zones.
+        assert zones == []
+
+
+# ---------------------------------------------------------------------------
+# Lowered MIN_AIRCRAFT (5 → 3)
+# ---------------------------------------------------------------------------
+
+
+class TestMinAircraftLowered:
+    def test_three_aircraft_cell_now_qualifies(self):
+        """Pre-fix MIN_AIRCRAFT=5 blocked sparse cells entirely. Post-fix
+        the bar is 3 aircraft per cell, which is realistic for the actual
+        jamming hotspots where traffic is thinner."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 33.1, "lng": 44.1, "nac_p": 3},
+            {"hex": "a2", "lat": 33.2, "lng": 44.2, "nac_p": 5},
+            {"hex": "a3", "lat": 33.3, "lng": 44.3, "nac_p": 7},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # total=3, degraded=3, adjusted=2, ratio=0.667 — fires under new
+        # rules, would have been blocked by MIN_AIRCRAFT=5 pre-fix.
+        assert len(zones) == 1
+
+    def test_two_aircraft_cell_still_blocked(self):
+        """We didn't lower the bar to 2 — that would create too much
+        single-transponder noise. Two aircraft per cell still doesn't
+        qualify."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 33.1, "lng": 44.1, "nac_p": 3},
+            {"hex": "a2", "lat": 33.2, "lng": 44.2, "nac_p": 3},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        assert zones == []
+
+
+# ---------------------------------------------------------------------------
+# Lowered MIN_RATIO (0.30 → 0.20)
+# ---------------------------------------------------------------------------
+
+
+class TestMinRatioLowered:
+    def test_ratio_between_old_and_new_threshold_fires(self):
+        """Construct a cell whose ratio sits in the (0.20, 0.30) window:
+        fires under the new bar, would have been blocked pre-fix."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        # 10 aircraft, 4 degraded → adjusted=3, ratio=3/10=0.30.
+        # Pre-fix threshold was > 0.30 strict — would NOT fire.
+        # Post-fix threshold is > 0.20 — fires.
+        feed = (
+            [{"hex": f"d{i}", "lat": 40.1, "lng": -100.1, "nac_p": 3} for i in range(4)]
+            + [{"hex": f"c{i}", "lat": 40.5, "lng": -100.5, "nac_p": 9} for i in range(6)]
+        )
+
+        zones = detect_gps_jamming_zones(feed)
+
+        assert len(zones) == 1
+        assert zones[0]["degraded"] == 4
+        assert zones[0]["total"] == 10
+        assert zones[0]["ratio"] == 0.30
+
+    def test_ratio_at_or_below_new_threshold_does_not_fire(self):
+        """Ratio of exactly 0.20 must NOT fire (strict ``>`` comparison)."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        # 15 aircraft, 4 degraded → adjusted=3, ratio=3/15=0.20. Strictly
+        # not greater than 0.20, so doesn't qualify.
+        feed = (
+            [{"hex": f"d{i}", "lat": 40.1, "lng": -100.1, "nac_p": 3} for i in range(4)]
+            + [{"hex": f"c{i}", "lat": 40.5, "lng": -100.5, "nac_p": 9} for i in range(11)]
+        )
+
+        zones = detect_gps_jamming_zones(feed)
+
+        assert zones == []
+
+
+# ---------------------------------------------------------------------------
+# Pre-existing noise cushion (-1) preserved
+# ---------------------------------------------------------------------------
+
+
+class TestNoiseCushionPreserved:
+    def test_single_quirky_transponder_doesnt_fire(self):
+        """One degraded aircraft in a healthy cell shouldn't fire even
+        under the relaxed thresholds. The ``-1`` adjustment in the
+        detector exists for this reason."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = (
+            [{"hex": "d1", "lat": 40.1, "lng": -100.1, "nac_p": 3}]
+            + [{"hex": f"c{i}", "lat": 40.5, "lng": -100.5, "nac_p": 9} for i in range(10)]
+        )
+
+        zones = detect_gps_jamming_zones(feed)
+
+        # total=11, degraded=1, adjusted=0 → cell short-circuits.
+        assert zones == []
+
+
+# ---------------------------------------------------------------------------
+# Constants pinned (catches accidental rollback)
+# ---------------------------------------------------------------------------
+
+
+class TestConstantsPinned:
+    def test_min_aircraft_is_three(self):
+        from services.constants import GPS_JAMMING_MIN_AIRCRAFT
+        assert GPS_JAMMING_MIN_AIRCRAFT == 3, (
+            "MIN_AIRCRAFT must be 3; raising it back to 5 brings back the "
+            "'jamming never shows' bug."
+        )
+
+    def test_min_ratio_is_0_20(self):
+        from services.constants import GPS_JAMMING_MIN_RATIO
+        assert GPS_JAMMING_MIN_RATIO == 0.20, (
+            "MIN_RATIO must be 0.20; raising it back to 0.30 brings back "
+            "the 'jamming never shows' bug."
+        )
+
+
+# ---------------------------------------------------------------------------
+# Overrides honored
+# ---------------------------------------------------------------------------
+
+
+class TestOverridesHonored:
+    def test_overrides_supersede_constants(self):
+        """The public signature accepts overrides so an operator can
+        re-tune at the call site (e.g. for a more aggressive setup in
+        an active conflict zone) without editing the module constants."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 40.1, "lng": -100.1, "nac_p": 3},
+            {"hex": "a2", "lat": 40.2, "lng": -100.2, "nac_p": 3},
+        ]
+
+        # With defaults (min_aircraft=3) this is blocked. With override=2 it fires.
+        assert detect_gps_jamming_zones(feed) == []
+        zones = detect_gps_jamming_zones(feed, min_aircraft=2)
+        assert len(zones) == 1
+
+
+# ---------------------------------------------------------------------------
+# lon vs lng compatibility
+# ---------------------------------------------------------------------------
+
+
+class TestLonLngCompat:
+    def test_lon_key_accepted(self):
+        """adsb.lol records arrive with ``lon`` (no g). The OpenSky merge
+        normalizes to ``lng`` but raw records flowing into the detector
+        may use either. Make sure both work."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+
+        feed = [
+            {"hex": "a1", "lat": 40.1, "lon": -100.1, "nac_p": 0},
+            {"hex": "a2", "lat": 40.2, "lon": -100.2, "nac_p": 0},
+            {"hex": "a3", "lat": 40.3, "lon": -100.3, "nac_p": 0},
+        ]
+
+        zones = detect_gps_jamming_zones(feed)
+
+        assert len(zones) == 1
+
+
+# ---------------------------------------------------------------------------
+# Empty / malformed inputs don't crash
+# ---------------------------------------------------------------------------
+
+
+class TestRobustness:
+    def test_empty_feed(self):
+        from services.fetchers.flights import detect_gps_jamming_zones
+        assert detect_gps_jamming_zones([]) == []
+
+    def test_none_feed(self):
+        """The wrapper at the call site passes ``raw_flights_snapshot``
+        which could in principle be None on a startup race. Handle it."""
+        from services.fetchers.flights import detect_gps_jamming_zones
+        assert detect_gps_jamming_zones(None) == []
+
+    def test_records_missing_position_skipped(self):
+        from services.fetchers.flights import detect_gps_jamming_zones
+        feed = [
+            {"hex": "noloc", "nac_p": 0},
+            {"hex": "nolat", "lng": -100.0, "nac_p": 0},
+            {"hex": "nolng", "lat": 40.0, "nac_p": 0},
+        ]
+        assert detect_gps_jamming_zones(feed) == []

backend/tests/test_openclaw_connect_info_reveal.py

@@ -0,0 +1,334 @@
+"""Issue #302 (tg12): OpenClaw connect-info HMAC secret disclosure.
+
+Before this change, ``GET /api/ai/connect-info?reveal=true`` returned the
+full HMAC secret in the response body on every modal open AND the same
+GET endpoint auto-bootstrapped (generated + persisted) the secret on a
+mere read. Even gated to ``require_local_operator``, that put the full
+secret into:
+
+  * browser visit history
+  * dev-tools network panel
+  * browser disk cache
+  * HAR exports
+  * screen captures / shoulder-surfing
+
+Every single time the OpenClaw Connect modal opened.
+
+After this change:
+
+  GET  /api/ai/connect-info            — always returns the MASKED
+                                          fingerprint. No ?reveal param.
+                                          No side effects (auto-bootstrap
+                                          gone).
+  POST /api/ai/connect-info/bootstrap  — mints+persists the secret if
+                                          missing. Idempotent. Never
+                                          returns the full secret.
+  POST /api/ai/connect-info/reveal     — returns the full secret with
+                                          strict Cache-Control: no-store
+                                          headers. POST so the body
+                                          doesn't land in URL history.
+  POST /api/ai/connect-info/regenerate — keeps the one-time-disclosure
+                                          for the new secret (regen IS a
+                                          deliberate destructive action).
+                                          Same no-store headers added.
+
+These tests pin every property.
+"""
+
+from __future__ import annotations
+
+import asyncio
+from unittest.mock import patch
+
+import pytest
+from httpx import ASGITransport, AsyncClient
+
+
+# ---------------------------------------------------------------------------
+# Loopback test client. ``require_local_operator`` resolves true for
+# request.client.host == "127.0.0.1"; FastAPI's TestClient sets it to
+# "testclient" which isn't on the allowlist. Use raw ASGITransport.
+# ---------------------------------------------------------------------------
+
+
+@pytest.fixture
+def loopback():
+    from main import app
+
+    class _Client:
+        def __init__(self, peer_ip: str = "127.0.0.1"):
+            self._loop = asyncio.new_event_loop()
+            self._transport = ASGITransport(app=app, client=(peer_ip, 12345))
+            self._base = f"http://{peer_ip}:8000"
+
+        def _do(self, method: str, url: str, **kw):
+            async def go():
+                async with AsyncClient(transport=self._transport, base_url=self._base) as ac:
+                    return await ac.request(method, url, **kw)
+            return self._loop.run_until_complete(go())
+
+        def get(self, url, **kw):  return self._do("GET", url, **kw)
+        def post(self, url, **kw): return self._do("POST", url, **kw)
+        def close(self): self._loop.close()
+
+    c = _Client()
+    yield c
+    c.close()
+
+
+@pytest.fixture
+def remote():
+    from main import app
+
+    class _Client:
+        def __init__(self):
+            self._loop = asyncio.new_event_loop()
+            self._transport = ASGITransport(app=app, client=("1.2.3.4", 12345))
+            self._base = "http://1.2.3.4:8000"
+
+        def _do(self, method: str, url: str, **kw):
+            async def go():
+                async with AsyncClient(transport=self._transport, base_url=self._base) as ac:
+                    return await ac.request(method, url, **kw)
+            return self._loop.run_until_complete(go())
+
+        def get(self, url, **kw):  return self._do("GET", url, **kw)
+        def post(self, url, **kw): return self._do("POST", url, **kw)
+        def close(self): self._loop.close()
+
+    c = _Client()
+    yield c
+    c.close()
+
+
+@pytest.fixture
+def stub_env(monkeypatch):
+    """Isolate connect-info tests from the dev's real backend .env.
+
+    Pydantic ``Settings()`` reads from ``.env`` file directly on
+    instantiation, so monkey-patching ``os.environ`` isn't sufficient
+    — the real ``OPENCLAW_HMAC_SECRET`` would leak through. Instead we
+    override ``get_settings()`` in the route module to return a fresh
+    ``Settings`` instance whose env values are driven entirely by an
+    in-test dict, AND we replace ``_write_env_value`` so writes update
+    that same dict instead of touching the developer's filesystem.
+
+    Yields the dict so individual tests can pre-seed values or assert
+    that writes happened.
+    """
+    import routers.ai_intel as ai_intel
+    import services.config as config
+
+    state: dict[str, str] = {}
+
+    class _FakeSettings:
+        @property
+        def OPENCLAW_HMAC_SECRET(self) -> str:
+            return state.get("OPENCLAW_HMAC_SECRET", "")
+
+        @property
+        def OPENCLAW_ACCESS_TIER(self) -> str:
+            return state.get("OPENCLAW_ACCESS_TIER", "restricted")
+
+    fake = _FakeSettings()
+
+    def _fake_get_settings():
+        return fake
+
+    # Route code calls ``get_settings.cache_clear()`` after writing the
+    # env. The production version is wrapped with ``@lru_cache``, so
+    # cache_clear exists. Attach a no-op shim here.
+    _fake_get_settings.cache_clear = lambda: None  # type: ignore[attr-defined]
+
+    monkeypatch.setattr(config, "get_settings", _fake_get_settings)
+
+    def _fake_write_env_value(key: str, value: str) -> None:
+        state[key] = value
+
+    monkeypatch.setattr(ai_intel, "_write_env_value", _fake_write_env_value)
+
+    yield state
+
+
+# ---------------------------------------------------------------------------
+# GET /api/ai/connect-info — always masked, no auto-bootstrap
+# ---------------------------------------------------------------------------
+
+
+class TestGetConnectInfoMasking:
+    def test_returns_masked_when_secret_set(self, loopback, stub_env):
+        secret = "abcdef" + "0" * 38 + "wxyz"
+        stub_env["OPENCLAW_HMAC_SECRET"] = secret
+
+        r = loopback.get("/api/ai/connect-info")
+        assert r.status_code == 200
+        body = r.json()
+        # Body must NOT carry the full secret value anywhere.
+        assert secret not in r.text, (
+            "GET /api/ai/connect-info MUST NOT include the full HMAC "
+            "secret. Response body contained the secret value."
+        )
+        assert body["hmac_secret_set"] is True
+        assert body["masked_hmac_secret"].startswith("abcdef")
+        assert body["masked_hmac_secret"].endswith("wxyz")
+        assert "•" in body["masked_hmac_secret"]
+        # Pre-fix field is gone.
+        assert "hmac_secret" not in body
+
+    def test_no_auto_bootstrap_when_secret_missing(self, loopback, stub_env):
+        """Side-effect-on-GET was the second half of issue #302. A GET
+        with no secret configured must NOT mint one — that should
+        require an explicit POST /bootstrap."""
+        r = loopback.get("/api/ai/connect-info")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["hmac_secret_set"] is False
+        assert body["masked_hmac_secret"] == ""
+        # The bootstrap_behavior block should advertise the new flow.
+        assert body["bootstrap_behavior"]["auto_generates_when_missing"] is False
+        # And no _write_env_value call happened.
+        assert "OPENCLAW_HMAC_SECRET" not in stub_env
+
+    def test_no_reveal_query_param(self, loopback, stub_env):
+        """Pre-fix, ?reveal=true would return the full secret. Post-fix
+        the param is silently ignored — the response is the same as
+        without it (still masked, no leak)."""
+        secret = "abcdef" + "0" * 38 + "wxyz"
+        stub_env["OPENCLAW_HMAC_SECRET"] = secret
+
+        r = loopback.get("/api/ai/connect-info?reveal=true")
+        assert r.status_code == 200
+        assert secret not in r.text, (
+            "?reveal=true must be a no-op on GET — the full secret "
+            "MUST NOT come back in the response body."
+        )
+
+
+# ---------------------------------------------------------------------------
+# POST /api/ai/connect-info/bootstrap
+# ---------------------------------------------------------------------------
+
+
+class TestBootstrap:
+    def test_mints_when_missing(self, loopback, stub_env):
+        r = loopback.post("/api/ai/connect-info/bootstrap")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["ok"] is True
+        assert body["generated"] is True
+        assert body["hmac_secret_set"] is True
+        # Bootstrap must NOT return the full secret in-line.
+        assert "hmac_secret" not in body or not body.get("hmac_secret")
+        assert "•" in body["masked_hmac_secret"]
+        # _write_env_value was actually called.
+        assert stub_env.get("OPENCLAW_HMAC_SECRET")
+        # The full value isn't echoed back in the response text either.
+        assert stub_env["OPENCLAW_HMAC_SECRET"] not in r.text
+
+    def test_idempotent_when_already_set(self, loopback, stub_env):
+        existing = "abcdef" + "0" * 38 + "wxyz"
+        stub_env["OPENCLAW_HMAC_SECRET"] = existing
+
+        r = loopback.post("/api/ai/connect-info/bootstrap")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["ok"] is True
+        assert body["generated"] is False
+        assert body["hmac_secret_set"] is True
+        # Existing secret untouched — value is still the seeded one.
+        assert stub_env["OPENCLAW_HMAC_SECRET"] == existing
+        # No full secret in the response.
+        assert existing not in r.text
+
+
+# ---------------------------------------------------------------------------
+# POST /api/ai/connect-info/reveal
+# ---------------------------------------------------------------------------
+
+
+class TestReveal:
+    def test_returns_full_secret_when_set(self, loopback, stub_env):
+        secret = "abcdef" + "0" * 38 + "wxyz"
+        stub_env["OPENCLAW_HMAC_SECRET"] = secret
+
+        r = loopback.post("/api/ai/connect-info/reveal")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["ok"] is True
+        assert body["hmac_secret"] == secret
+
+    def test_strict_cache_control_headers(self, loopback, stub_env):
+        """The whole point of POST /reveal vs GET ?reveal=true is that
+        the response carries headers that prevent every cache layer
+        from persisting the secret."""
+        secret = "abcdef" + "0" * 38 + "wxyz"
+        stub_env["OPENCLAW_HMAC_SECRET"] = secret
+
+        r = loopback.post("/api/ai/connect-info/reveal")
+        cc = r.headers.get("cache-control", "")
+        assert "no-store" in cc, (
+            f"reveal MUST set Cache-Control: no-store — got {cc!r}"
+        )
+        assert "no-cache" in cc
+        # Pragma + Expires as well for HTTP/1.0 caches.
+        assert r.headers.get("pragma", "").lower() == "no-cache"
+        assert r.headers.get("expires") == "0"
+
+    def test_404_when_no_secret_configured(self, loopback, stub_env):
+        r = loopback.post("/api/ai/connect-info/reveal")
+        assert r.status_code == 404
+        # Hint should point at the bootstrap endpoint, not just say "404".
+        detail = r.json().get("detail", "")
+        assert "/bootstrap" in detail or "bootstrap" in detail.lower()
+
+
+# ---------------------------------------------------------------------------
+# POST /api/ai/connect-info/regenerate — still returns the new secret
+# inline (deliberate destructive action), but with no-store headers.
+# ---------------------------------------------------------------------------
+
+
+class TestRegenerate:
+    def test_returns_new_secret_with_no_store_headers(self, loopback, stub_env):
+        # Seed an existing secret so we can prove it changes.
+        old = "oldold" + "0" * 38 + "1234"
+        stub_env["OPENCLAW_HMAC_SECRET"] = old
+
+        r = loopback.post("/api/ai/connect-info/regenerate")
+        assert r.status_code == 200
+        body = r.json()
+        assert body["ok"] is True
+        assert body["hmac_secret"]
+        assert body["hmac_secret"] != old
+        # no-store headers MUST be present so the new secret doesn't
+        # land in browser disk cache after the regenerate click.
+        cc = r.headers.get("cache-control", "")
+        assert "no-store" in cc and "no-cache" in cc
+        assert r.headers.get("pragma", "").lower() == "no-cache"
+
+
+# ---------------------------------------------------------------------------
+# Auth-gate regression — every endpoint still rejects anonymous remote
+# callers. This is the property we already enforce for the rest of the
+# operator-only surface; adding the three new endpoints to the audit
+# coverage prevents a future refactor from dropping the dependency.
+# ---------------------------------------------------------------------------
+
+
+class TestAnonymousRejection:
+    @pytest.mark.parametrize(
+        "method,path,body",
+        [
+            ("get",  "/api/ai/connect-info",            None),
+            ("post", "/api/ai/connect-info/bootstrap",  None),
+            ("post", "/api/ai/connect-info/reveal",     None),
+            ("post", "/api/ai/connect-info/regenerate", None),
+        ],
+    )
+    def test_remote_rejected(self, remote, method, path, body):
+        fn = getattr(remote, method)
+        r = fn(path, json=body) if body is not None else fn(path)
+        assert r.status_code == 403, (
+            f"{method.upper()} {path} must reject anonymous remote callers; "
+            f"got {r.status_code}"
+        )

backend/tests/test_uap_hf_fallback_cutoff.py

@@ -0,0 +1,252 @@
+"""HF NUFORC fallback honors the rolling cutoff window.
+
+Background
+----------
+The UAP sightings layer is sourced primarily from a live scrape of
+nuforc.org. When that fails (Cloudflare 403, curl disabled on Windows,
+wdtNonce regex stale, etc.) the code falls back to a static CSV mirror
+hosted on Hugging Face at ``kcimc/NUFORC/nuforc_str.csv``.
+
+The HF mirror is maintained by a third party and refreshed sporadically.
+Pre-fix, the fallback parsed every row, sorted by ``occurred`` descending,
+and took the top 250 — **with no date cutoff**. When the HF mirror is
+stale (its "newest" rows are ~2-3 years old), users saw a map full of
+2022-2023 sightings labeled as the "last 60 days" layer.
+
+These tests pin the new behavior:
+
+* Rows older than ``_NUFORC_RECENT_DAYS`` are dropped before the take-top-N.
+* If the HF mirror has nothing in the window, the fallback returns ``[]``
+  and logs ERROR (don't silently serve stale data).
+* ``fetch_uap_sightings`` records the failure when BOTH paths fail, so
+  the layer shows as broken in the health registry instead of "fresh".
+"""
+
+from __future__ import annotations
+
+import logging
+from datetime import datetime as real_datetime
+
+
+class _FixedDateTime(real_datetime):
+    """A datetime whose utcnow() returns a pinned value, for deterministic
+    cutoff math. Subclasses real datetime so existing operations still work."""
+
+    @classmethod
+    def utcnow(cls):
+        return cls(2026, 5, 1, 12, 0, 0)
+
+
+class _StubResponse:
+    status_code = 200
+
+    def __init__(self, text: str):
+        self.text = text
+
+
+def _stub_geocode_cache(*_args, **_kwargs):
+    """Pre-populated location cache so the fallback doesn't try to hit
+    Photon during the test."""
+    return {
+        "Denver, CO, USA": [39.7392, -104.9903],
+        "Seattle, WA, USA": [47.6062, -122.3321],
+        "Phoenix, AZ, USA": [33.4484, -112.0740],
+    }
+
+
+def test_hf_fallback_drops_rows_older_than_60_days(monkeypatch):
+    """Pre-fix: a row from 2023 would make it into the layer if it was
+    among the newest 250 in the HF mirror. Post-fix: it's filtered out
+    before we even count to 250."""
+    from services.fetchers import earth_observation as eo
+
+    # 2026-05-01 - 60 days = 2026-03-02. So 2026-03-01 is one day too old.
+    csv_text = (
+        "Sighting,Occurred,Location,Shape,Duration,Posted,Summary\n"
+        '1,2026-04-15 21:00:00 Local,"Denver, CO, USA",Triangle,5 minutes,2026-04-16,"In-window sighting"\n'
+        '2,2023-06-01 21:00:00 Local,"Seattle, WA, USA",Light,30 seconds,2023-06-02,"Three years old"\n'
+        '3,2022-01-15 20:00:00 Local,"Phoenix, AZ, USA",Disk,2 minutes,2022-01-16,"Even older"\n'
+    )
+
+    monkeypatch.setattr(eo, "datetime", _FixedDateTime)
+    monkeypatch.setattr(eo, "fetch_with_curl", lambda *a, **kw: _StubResponse(csv_text))
+    monkeypatch.setattr(eo, "_load_nuforc_location_cache", _stub_geocode_cache)
+    monkeypatch.setattr(eo, "_save_nuforc_location_cache", lambda cache: None)
+    # If the cutoff is missing, the geocoder may still get called for the
+    # 2022/2023 rows. We assert geocoder is NEVER invoked for stale rows.
+    geocode_calls: list[str] = []
+
+    def _geocode_spy(location, city, state, country=""):
+        geocode_calls.append(location)
+        return None  # already in cache, shouldn't be hit anyway
+
+    monkeypatch.setattr(eo, "_geocode_uap_location", _geocode_spy)
+
+    sightings = eo._build_uap_sightings_from_hf_mirror()
+
+    ids = [s["id"] for s in sightings]
+    assert ids == ["NUFORC-1"], f"only the 2026 row should survive: got {ids}"
+    # Stale rows must not have been geocoded — they should be dropped
+    # before the geocoding loop is reached.
+    assert geocode_calls == []
+
+
+def test_hf_fallback_returns_empty_when_mirror_is_fully_stale(monkeypatch, caplog):
+    """The smoking-gun case: the HF mirror is so stale that NO rows are
+    within the rolling window. Pre-fix returned 250 ancient rows. Post-fix
+    returns ``[]`` and logs ERROR so the operator knows the layer is dead."""
+    from services.fetchers import earth_observation as eo
+
+    csv_text = (
+        "Sighting,Occurred,Location,Shape,Duration,Posted,Summary\n"
+        '1,2023-04-15 21:00:00 Local,"Denver, CO, USA",Triangle,5 minutes,2023-04-16,"Old"\n'
+        '2,2022-06-01 21:00:00 Local,"Seattle, WA, USA",Light,30 seconds,2022-06-02,"Older"\n'
+        '3,2021-01-15 20:00:00 Local,"Phoenix, AZ, USA",Disk,2 minutes,2021-01-16,"Ancient"\n'
+    )
+
+    monkeypatch.setattr(eo, "datetime", _FixedDateTime)
+    monkeypatch.setattr(eo, "fetch_with_curl", lambda *a, **kw: _StubResponse(csv_text))
+    monkeypatch.setattr(eo, "_load_nuforc_location_cache", _stub_geocode_cache)
+    monkeypatch.setattr(eo, "_save_nuforc_location_cache", lambda cache: None)
+    monkeypatch.setattr(eo, "_geocode_uap_location", lambda *a, **kw: None)
+
+    with caplog.at_level(logging.ERROR, logger="services.fetchers.earth_observation"):
+        sightings = eo._build_uap_sightings_from_hf_mirror()
+
+    assert sightings == []
+    # The error log should mention how many stale rows were dropped so the
+    # operator can tell the mirror is the problem (not "we got 0 rows" which
+    # could also mean the download failed).
+    relevant = [r for r in caplog.records if "HF fallback yielded 0 rows" in r.getMessage()]
+    assert relevant, "expected loud ERROR when HF mirror is fully stale"
+    # The message should report the count of dropped stale rows.
+    assert any("dropped 3" in r.getMessage() for r in relevant)
+
+
+def test_hf_fallback_still_returns_data_when_some_rows_are_in_window(monkeypatch):
+    """Mixed-age mirror: some rows in the window, some not. The fallback
+    should return only the in-window rows and not log the doomsday ERROR."""
+    from services.fetchers import earth_observation as eo
+
+    csv_text = (
+        "Sighting,Occurred,Location,Shape,Duration,Posted,Summary\n"
+        '1,2026-04-15 21:00:00 Local,"Denver, CO, USA",Triangle,5 minutes,2026-04-16,"Fresh"\n'
+        '2,2026-04-10 21:00:00 Local,"Seattle, WA, USA",Light,30 seconds,2026-04-10,"Also fresh"\n'
+        '3,2020-01-15 20:00:00 Local,"Phoenix, AZ, USA",Disk,2 minutes,2020-01-16,"Ancient"\n'
+    )
+
+    monkeypatch.setattr(eo, "datetime", _FixedDateTime)
+    monkeypatch.setattr(eo, "fetch_with_curl", lambda *a, **kw: _StubResponse(csv_text))
+    monkeypatch.setattr(eo, "_load_nuforc_location_cache", _stub_geocode_cache)
+    monkeypatch.setattr(eo, "_save_nuforc_location_cache", lambda cache: None)
+    monkeypatch.setattr(eo, "_geocode_uap_location", lambda *a, **kw: None)
+
+    sightings = eo._build_uap_sightings_from_hf_mirror()
+
+    ids = sorted(s["id"] for s in sightings)
+    assert ids == ["NUFORC-1", "NUFORC-2"], f"only in-window rows should appear: got {ids}"
+
+
+def test_fetch_uap_sightings_marks_failure_when_both_paths_empty(monkeypatch, caplog):
+    """When the live path raises AND the HF fallback returns empty,
+    ``fetch_uap_sightings`` must:
+      * NOT mark the layer fresh (pre-fix bug: it did, so the layer
+        showed as healthy-but-empty for days)
+      * call ``assert_canary("uap_sightings", 0)`` so the health
+        registry surfaces the broken layer
+      * log an ERROR with the live-path exception for debugging
+    """
+    from services.fetchers import earth_observation as eo
+    from services.fetchers import _store
+
+    monkeypatch.setattr(_store, "is_any_active", lambda layer: True)
+    monkeypatch.setattr(eo, "_load_nuforc_sightings_cache", lambda force_refresh=False: None)
+
+    def _boom():
+        raise RuntimeError("NUFORC live: zero rows pulled across 3 months")
+
+    monkeypatch.setattr(eo, "_build_recent_uap_sightings", _boom)
+    monkeypatch.setattr(eo, "_build_uap_sightings_from_hf_mirror", lambda: [])
+
+    marked: list[str] = []
+    monkeypatch.setattr(eo, "_mark_fresh", lambda *keys: marked.extend(keys))
+
+    canary_calls: list[tuple[str, int]] = []
+    import services.slo as slo
+    monkeypatch.setattr(
+        slo, "assert_canary", lambda key, value: canary_calls.append((key, int(value)))
+    )
+
+    with caplog.at_level(logging.ERROR, logger="services.fetchers.earth_observation"):
+        eo.fetch_uap_sightings()
+
+    assert marked == [], "broken layer must NOT be marked fresh"
+    assert canary_calls == [("uap_sightings", 0)], (
+        f"expected canary trip when both paths fail; got {canary_calls}"
+    )
+    # The live error message should propagate into the error log so the
+    # operator can tell live failed AND fallback was empty (not the other
+    # way around).
+    assert any(
+        "both live NUFORC and HF fallback" in r.getMessage()
+        for r in caplog.records
+    )
+
+
+def test_fetch_uap_sightings_succeeds_when_fallback_returns_data(monkeypatch):
+    """Positive path: live fails, fallback returns rows. The layer is
+    populated and marked fresh; assert_canary is NOT tripped (we only
+    trip the canary when the layer has zero data)."""
+    from services.fetchers import earth_observation as eo
+    from services.fetchers import _store
+
+    monkeypatch.setattr(_store, "is_any_active", lambda layer: True)
+    monkeypatch.setattr(eo, "_load_nuforc_sightings_cache", lambda force_refresh=False: None)
+    monkeypatch.setattr(
+        eo, "_build_recent_uap_sightings", lambda: (_ for _ in ()).throw(RuntimeError("live down"))
+    )
+
+    fallback_rows = [{"id": "NUFORC-fb-1", "date_time": "2026-04-20", "lat": 0.0, "lng": 0.0}]
+    monkeypatch.setattr(eo, "_build_uap_sightings_from_hf_mirror", lambda: fallback_rows)
+    monkeypatch.setattr(eo, "_save_nuforc_sightings_cache", lambda s: None)
+
+    marked: list[str] = []
+    monkeypatch.setattr(eo, "_mark_fresh", lambda *keys: marked.extend(keys))
+
+    canary_calls: list[tuple[str, int]] = []
+    import services.slo as slo
+    monkeypatch.setattr(
+        slo, "assert_canary", lambda key, value: canary_calls.append((key, int(value)))
+    )
+
+    eo.fetch_uap_sightings()
+
+    assert marked == ["uap_sightings"]
+    assert canary_calls == [], "canary should not trip when fallback supplies data"
+
+
+def test_uap_scheduler_runs_weekly_not_daily():
+    """The cron job for the UAP layer must be configured for Mondays at
+    12:00 UTC, not daily. Daily was the pre-fix default; weekly matches
+    the layer's stated cadence (a rolling 60-day digest) and keeps load
+    on nuforc.org light."""
+    from services import data_fetcher
+
+    src = data_fetcher.__file__
+    with open(src, "r", encoding="utf-8") as f:
+        text = f.read()
+
+    # Anchor on the scheduler block by id, then assert the cron triggers.
+    assert "uap_sightings_weekly" in text, (
+        "scheduler id should be uap_sightings_weekly (was uap_sightings_daily pre-fix)"
+    )
+    # The day_of_week directive is the difference between daily and weekly.
+    # If somebody flips it back to daily, this fires.
+    weekly_block = text.split("uap_sightings_weekly", 1)[0]
+    # Walk backwards for the matching add_job call.
+    add_job_idx = weekly_block.rfind("add_job(")
+    assert add_job_idx >= 0, "could not locate add_job block for UAP scheduler"
+    job_block = text[add_job_idx : text.find(")", text.index("uap_sightings_weekly")) + 1]
+    assert 'day_of_week="mon"' in job_block, (
+        f"expected day_of_week='mon' in UAP scheduler block:\n{job_block}"
+    )

frontend/src/app/page.tsx

@@ -39,6 +39,7 @@ import { useFeedHealth } from '@/hooks/useFeedHealth';
 import { useKeyboardShortcuts } from '@/hooks/useKeyboardShortcuts';
 import KeyboardShortcutsOverlay from '@/components/KeyboardShortcutsOverlay';
 import AlertToast from '@/components/AlertToast';
+import AisUpstreamBanner from '@/components/AisUpstreamBanner';
 import { useAlertToasts } from '@/hooks/useAlertToasts';
 import { useWatchlist } from '@/hooks/useWatchlist';
 import WatchlistWidget from '@/components/WatchlistWidget';
@@ -933,6 +934,11 @@ export default function Dashboard() {
           onFlyTo={handleFlyTo}
         />
 
+        {/* AIS UPSTREAM OUTAGE BANNER — renders only when AIS is configured
+            but the WebSocket upstream is unreachable. Tells users the empty
+            ocean isn't their fault. */}
+        <AisUpstreamBanner />
+
         {/* ONBOARDING MODAL */}
         {showOnboarding && (
           <OnboardingModal

frontend/src/components/AIIntelPanel.tsx

@@ -357,8 +357,15 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
   const [riskAccepted, setRiskAccepted] = React.useState(false);
   const [accessTier, setAccessTier] = React.useState<'restricted' | 'full'>('restricted');
   const [connectionMode, setConnectionMode] = React.useState<'local' | 'remote'>('local');
+  // hmacSecret holds the FULL secret once the operator has clicked
+  // Reveal (or after a regenerate). maskedHmacSecret is the safe-to-show
+  // fingerprint returned by GET /api/ai/connect-info and is loaded on
+  // mount. The two are independent state slots so a stale full secret
+  // can never leak back into the UI after a regenerate.
   const [hmacSecret, setHmacSecret] = React.useState('');
+  const [maskedHmacSecret, setMaskedHmacSecret] = React.useState('');
   const [hmacLoading, setHmacLoading] = React.useState(false);
+  const [revealing, setRevealing] = React.useState(false);
   const [tierSaving, setTierSaving] = React.useState(false);
   const [showAdvanced, setShowAdvanced] = React.useState(false);
   const [showResetConfirm, setShowResetConfirm] = React.useState(false);
@@ -381,16 +388,40 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
   const [torError, setTorError] = React.useState('');
   const [torOnion, setTorOnion] = React.useState('');
 
-  // Fetch connect-info + node status on mount
+  // Issue #302 (tg12): the full HMAC secret no longer travels through
+  // GET /api/ai/connect-info on every modal open. The flow is now:
+  //
+  //   1. GET /api/ai/connect-info — always returns the masked fingerprint
+  //      (first6 + bullets + last4). `hmacSecret` stays empty until the
+  //      operator clicks the Reveal (eye) button below.
+  //   2. POST /api/ai/connect-info/bootstrap — fires once on mount if the
+  //      backend reports `hmac_secret_set: false`. Idempotent and never
+  //      returns the secret in the response.
+  //   3. POST /api/ai/connect-info/reveal — fires when the operator clicks
+  //      Reveal or Copy without the secret yet loaded. Returns the full
+  //      secret with strict `Cache-Control: no-store` so it doesn't land
+  //      in browser caches or HAR exports.
   React.useEffect(() => {
     (async () => {
       try {
         setHmacLoading(true);
-        const res = await fetch(`${API_BASE}/api/ai/connect-info?reveal=true`);
-        if (res.ok) {
-          const data = await res.json();
-          setHmacSecret(data.hmac_secret || '');
-          setAccessTier(data.access_tier === 'full' ? 'full' : 'restricted');
+        const res = await fetch(`${API_BASE}/api/ai/connect-info`);
+        if (!res.ok) return;
+        const data = await res.json();
+        setMaskedHmacSecret(data.masked_hmac_secret || '');
+        setAccessTier(data.access_tier === 'full' ? 'full' : 'restricted');
+
+        // Transparent first-use bootstrap. Mirrors the pre-#302 UX of
+        // "open modal → secret exists" without the GET side-effect.
+        if (!data.hmac_secret_set) {
+          const bootRes = await fetch(
+            `${API_BASE}/api/ai/connect-info/bootstrap`,
+            { method: 'POST' },
+          );
+          if (bootRes.ok) {
+            const bootData = await bootRes.json();
+            setMaskedHmacSecret(bootData.masked_hmac_secret || '');
+          }
         }
       } catch { /* ignore */ }
       finally { setHmacLoading(false); }
@@ -477,8 +508,17 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
       const res = await fetch(`${API_BASE}/api/settings/agent/reset-all`, { method: 'POST' });
       const data = await res.json();
       if (data.ok) {
-        // Update local state with new credentials
-        if (data.new_hmac_secret) setHmacSecret(data.new_hmac_secret);
+        // Update local state with new credentials. reset-all returns
+        // the new HMAC secret in-band (same one-time-disclosure rule
+        // as /regenerate — a deliberate destructive action). Refresh
+        // both slots so the masked display stays in sync.
+        if (data.new_hmac_secret) {
+          setHmacSecret(data.new_hmac_secret);
+          const s = String(data.new_hmac_secret);
+          setMaskedHmacSecret(
+            s.length > 10 ? s.slice(0, 6) + '•'.repeat(8) + s.slice(-4) : '•'.repeat(16),
+          );
+        }
         if (data.new_onion) {
           setTorOnion(data.new_onion);
           setRemoteUrl(data.new_onion);
@@ -502,13 +542,41 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
     finally { setTierSaving(false); }
   };
 
+  // Issue #302: POST /reveal returns the full secret with strict
+  // no-store headers. Lazily fetched — never on mount. Returns the
+  // secret string so callers can copy it immediately without waiting
+  // for React state propagation.
+  const revealHmacSecret = async (): Promise<string> => {
+    if (hmacSecret) return hmacSecret;
+    setRevealing(true);
+    try {
+      const res = await fetch(`${API_BASE}/api/ai/connect-info/reveal`, {
+        method: 'POST',
+      });
+      if (!res.ok) return '';
+      const data = await res.json();
+      const secret = String(data.hmac_secret || '');
+      setHmacSecret(secret);
+      return secret;
+    } catch {
+      return '';
+    } finally {
+      setRevealing(false);
+    }
+  };
+
   const handleRegenerate = async () => {
     setRegenerating(true);
     try {
       const res = await fetch(`${API_BASE}/api/ai/connect-info/regenerate`, { method: 'POST' });
       if (res.ok) {
         const data = await res.json();
+        // Regenerate is a deliberate destructive action — operator needs
+        // to see the new secret once to update their OpenClaw config.
+        // Both the full and masked forms refresh in one shot.
         setHmacSecret(data.hmac_secret || '');
+        setMaskedHmacSecret(data.masked_hmac_secret || '');
+        setShowSecret(true);
       }
     } catch { /* ignore */ }
     finally { setRegenerating(false); }
@@ -543,9 +611,17 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
     finally { setNodeToggling(false); }
   };
 
-  const maskedSecret = hmacSecret
-    ? hmacSecret.slice(0, 6) + '\u2022'.repeat(8) + hmacSecret.slice(-4)
-    : '\u2022'.repeat(16);
+  // Issue #302: prefer the server-supplied fingerprint
+  // (maskedHmacSecret) \u2014 it's filled on mount via the (no-secret) GET.
+  // If the operator has clicked Reveal, fall through to deriving the
+  // mask from the in-memory full secret so we keep the same shape
+  // (first6 + bullets + last4) regardless of source. Final fallback
+  // (no secret loaded yet) is a generic bullet string.
+  const maskedSecret =
+    maskedHmacSecret ||
+    (hmacSecret
+      ? hmacSecret.slice(0, 6) + '\u2022'.repeat(8) + hmacSecret.slice(-4)
+      : '\u2022'.repeat(16));
 
   // Resolve the endpoint URL
   const resolvedUrl = connectionMode === 'local'
@@ -672,10 +748,15 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
     return lines.join('\n');
   };
   const displaySnippet = buildSnippet(maskedSecret);
-  const copySnippet = buildSnippet(hmacSecret);
 
-  const handleCopySnippet = () => {
-    navigator.clipboard.writeText(copySnippet);
+  // Issue #302: the copy snippet needs the FULL secret. Pre-#302 we kept
+  // it in memory from the GET-with-reveal load; now we lazy-fetch via
+  // POST /reveal only when the operator actually clicks Copy. If they
+  // already revealed, the in-memory value is reused (no extra request).
+  const handleCopySnippet = async () => {
+    const secret = hmacSecret || (await revealHmacSecret());
+    if (!secret) return;
+    navigator.clipboard.writeText(buildSnippet(secret));
     setSnippetCopied(true);
     setTimeout(() => setSnippetCopied(false), 2000);
   };
@@ -913,18 +994,38 @@ function ConnectModalBody({ apiEndpoint, handleCopy, copied }: ConnectModalBodyP
                   </div>
                   <div className="flex items-center gap-2">
                     <code className="flex-1 bg-black/60 border border-violet-800/40 px-3 py-2 text-xs font-mono text-violet-300 overflow-hidden text-ellipsis">
-                      {showSecret ? hmacSecret : maskedSecret}
+                      {/* Issue #302: when the operator hasn't clicked
+                          Reveal yet, hmacSecret is empty and we fall
+                          back to maskedHmacSecret (the safe fingerprint
+                          returned by GET /api/ai/connect-info). */}
+                      {showSecret && hmacSecret ? hmacSecret : (maskedHmacSecret || maskedSecret)}
                     </code>
                     <button
-                      onClick={() => setShowSecret(!showSecret)}
-                      className="p-2 bg-violet-600/20 border border-violet-500/40 text-violet-400 hover:bg-violet-600/40 transition-colors shrink-0"
+                      onClick={async () => {
+                        if (showSecret) {
+                          setShowSecret(false);
+                          return;
+                        }
+                        // Need the full secret in state before showing it.
+                        const secret = await revealHmacSecret();
+                        if (secret) setShowSecret(true);
+                      }}
+                      disabled={revealing}
+                      className="p-2 bg-violet-600/20 border border-violet-500/40 text-violet-400 hover:bg-violet-600/40 transition-colors shrink-0 disabled:opacity-50"
                       title={showSecret ? 'Hide' : 'Reveal'}
                     >
                       {showSecret ? <EyeOff size={14} /> : <Eye size={14} />}
                     </button>
                     <button
-                      onClick={() => handleCopy(hmacSecret)}
-                      className="p-2 bg-violet-600/20 border border-violet-500/40 text-violet-400 hover:bg-violet-600/40 transition-colors shrink-0"
+                      onClick={async () => {
+                        // Copy needs the full secret. Fetch it lazily if
+                        // the operator hasn't clicked Reveal yet — no
+                        // point making them reveal first just to copy.
+                        const secret = hmacSecret || (await revealHmacSecret());
+                        if (secret) handleCopy(secret);
+                      }}
+                      disabled={revealing}
+                      className="p-2 bg-violet-600/20 border border-violet-500/40 text-violet-400 hover:bg-violet-600/40 transition-colors shrink-0 disabled:opacity-50"
                       title="Copy key"
                     >
                       {copied ? <Check size={14} /> : <Copy size={14} />}

frontend/src/components/AisUpstreamBanner.tsx

@@ -0,0 +1,61 @@
+/**
+ * AisUpstreamBanner — visible notice that AIS ship data is unavailable
+ * because the upstream provider (AISStream) is offline.
+ *
+ * Renders nothing when AIS is healthy or when AIS isn't configured at all.
+ * Mounted at the app shell level so users see it before they wonder why
+ * the ocean looks empty.
+ */
+import { useState } from 'react';
+import { useAisUpstreamHealth } from '@/hooks/useAisUpstreamHealth';
+
+export function AisUpstreamBanner() {
+  const health = useAisUpstreamHealth();
+  const [dismissed, setDismissed] = useState(false);
+
+  if (!health || !health.aisEnabled || health.connected || dismissed) {
+    return null;
+  }
+
+  // Format the staleness for the operator. ``null`` means we never received
+  // anything since startup; otherwise show minutes if > 60s.
+  let stalenessLabel = 'never received';
+  if (health.lastMsgAgeSeconds != null) {
+    const minutes = Math.floor(health.lastMsgAgeSeconds / 60);
+    if (minutes >= 1) {
+      stalenessLabel = `last update ${minutes} min ago`;
+    } else {
+      stalenessLabel = `last update ${health.lastMsgAgeSeconds}s ago`;
+    }
+  }
+
+  return (
+    <div
+      role="status"
+      aria-live="polite"
+      className="pointer-events-auto fixed top-3 left-1/2 z-[100] -translate-x-1/2 max-w-[640px] rounded-md border border-amber-500/60 bg-amber-900/85 px-4 py-2 text-sm text-amber-50 shadow-lg backdrop-blur"
+    >
+      <div className="flex items-start gap-3">
+        <span aria-hidden className="mt-0.5 text-amber-300">⚠</span>
+        <div className="flex-1">
+          <div className="font-semibold">Ship data temporarily unavailable</div>
+          <div className="text-xs opacity-90">
+            AISStream upstream is offline ({stalenessLabel}). The map will
+            refill once their service comes back online — nothing is wrong
+            with your install.
+          </div>
+        </div>
+        <button
+          type="button"
+          onClick={() => setDismissed(true)}
+          aria-label="Dismiss"
+          className="text-amber-200 hover:text-white"
+        >
+          ✕
+        </button>
+      </div>
+    </div>
+  );
+}
+
+export default AisUpstreamBanner;

frontend/src/components/NewsFeed.tsx

@@ -249,34 +249,70 @@ const VESSEL_TYPE_WIKI: Record<string, string> = {
 
 type FlightTrailPoint = { lat?: number; lng?: number; alt?: number; ts?: number } | number[];
 
+function formatObservedDuration(seconds: number): string {
+    // Compact "1h 14m" / "23m" / "45s" — matches the density of the rest
+    // of the flight tooltip. < 60s is shown as "<1m" so the user knows
+    // we've JUST started observing this hex (cumulative will still be 0).
+    if (!Number.isFinite(seconds) || seconds <= 0) return '<1m';
+    if (seconds < 60) return '<1m';
+    const totalMinutes = Math.floor(seconds / 60);
+    const hours = Math.floor(totalMinutes / 60);
+    const minutes = totalMinutes % 60;
+    if (hours > 0) return `${hours}h ${minutes}m`;
+    return `${minutes}m`;
+}
+
 function EmissionsEstimateBlock({ flight }: { flight: any }) {
     const emissions = flight?.emissions;
-    const context = emissions ? 'Model-based cruise estimate' : null;
+    // Cumulative fuel/CO2 since the backend first saw this hex this
+    // flight session. Prefer these big numbers — the user explicitly
+    // wanted "the actual fuel that has been burned", not the rate.
+    // Rates are still shown below as smaller context.
+    const observedSec = Number(emissions?.observed_seconds ?? 0);
+    const fuelBurned = Number(emissions?.fuel_gallons_burned ?? 0);
+    const co2Emitted = Number(emissions?.co2_kg_emitted ?? 0);
+    const haveCumulative = emissions && observedSec > 0;
 
     return (
         <div className="border-b border-[var(--border-primary)] pb-2">
             <span className="text-[var(--text-muted)] text-[10px] block mb-1.5">EMISSIONS ESTIMATE</span>
             <div className="flex gap-3">
                 <div className="flex-1 bg-[var(--bg-primary)]/50 border border-[var(--border-primary)] px-2 py-1.5">
-                    <div className="text-[11px] text-[var(--text-muted)] tracking-widest">FUEL RATE</div>
-                    <div className="text-xs font-bold text-orange-400">
-                        {emissions ? (
-                            <>{emissions.fuel_gph} <span className="text-[11px] text-[var(--text-muted)] font-normal">GPH</span></>
+                    <div className="text-[11px] text-[var(--text-muted)] tracking-widest">FUEL BURNED</div>
+                    <div className="text-sm font-bold text-orange-400">
+                        {haveCumulative ? (
+                            <>{fuelBurned.toLocaleString(undefined, { maximumFractionDigits: 1 })} <span className="text-[11px] text-[var(--text-muted)] font-normal">gal</span></>
+                        ) : emissions ? (
+                            <span className="text-[var(--text-muted)] font-normal text-xs">—</span>
                         ) : 'UNKNOWN'}
                     </div>
+                    {emissions && (
+                        <div className="text-[10px] text-[var(--text-muted)] mt-0.5">
+                            @ {emissions.fuel_gph} gph
+                        </div>
+                    )}
                 </div>
                 <div className="flex-1 bg-[var(--bg-primary)]/50 border border-[var(--border-primary)] px-2 py-1.5">
-                    <div className="text-[11px] text-[var(--text-muted)] tracking-widest">CO2 RATE</div>
-                    <div className="text-xs font-bold text-red-400">
-                        {emissions ? (
-                            <>{emissions.co2_kg_per_hour.toLocaleString()} <span className="text-[11px] text-[var(--text-muted)] font-normal">KG/HR</span></>
+                    <div className="text-[11px] text-[var(--text-muted)] tracking-widest">CO2 EMITTED</div>
+                    <div className="text-sm font-bold text-red-400">
+                        {haveCumulative ? (
+                            <>{co2Emitted.toLocaleString(undefined, { maximumFractionDigits: 1 })} <span className="text-[11px] text-[var(--text-muted)] font-normal">kg</span></>
+                        ) : emissions ? (
+                            <span className="text-[var(--text-muted)] font-normal text-xs">—</span>
                         ) : 'UNKNOWN'}
                     </div>
+                    {emissions && (
+                        <div className="text-[10px] text-[var(--text-muted)] mt-0.5">
+                            @ {emissions.co2_kg_per_hour.toLocaleString()} kg/hr
+                        </div>
+                    )}
                 </div>
             </div>
-            {context && (
+            {emissions && (
                 <div className="mt-1.5 text-[10px] text-[var(--text-muted)] leading-relaxed">
-                    {context}
+                    {haveCumulative
+                        ? `Observed in flight for ${formatObservedDuration(observedSec)} · model-based cruise estimate`
+                        : 'Just observed · totals will appear on next refresh'}
                 </div>
             )}
         </div>

frontend/src/components/OnboardingModal.tsx

@@ -140,17 +140,51 @@ const OnboardingModal = React.memo(function OnboardingModal({
   ].join('\n');
   const remoteAgentNeedsTor = agentMode === 'remote' && !torAddress;
 
+  // Issue #302 (tg12): the full HMAC secret no longer comes back from
+  // GET /api/ai/connect-info. We fetch metadata + the masked fingerprint
+  // first; if the operator has explicitly asked to see the key (the
+  // ``reveal`` flag), we follow up with POST /api/ai/connect-info/reveal
+  // (after a transparent POST /bootstrap if the secret hasn't been
+  // minted yet) which carries the secret with strict no-store headers.
   const fetchAgentConnectInfo = async (reveal = true) => {
     setAgentLoading(true);
     setAgentMsg(null);
     try {
-      const res = await fetch(`/api/ai/connect-info?reveal=${reveal ? 'true' : 'false'}`);
-      const data = await res.json().catch(() => ({}));
-      if (!res.ok || data?.ok === false) {
-        throw new Error(data?.detail || 'Could not prepare agent credentials.');
+      // 1) GET metadata + masked fingerprint.
+      const metaRes = await fetch('/api/ai/connect-info');
+      const metaData = await metaRes.json().catch(() => ({}));
+      if (!metaRes.ok || metaData?.ok === false) {
+        throw new Error(metaData?.detail || 'Could not prepare agent credentials.');
+      }
+      setAgentTier(metaData.access_tier === 'full' ? 'full' : 'restricted');
+
+      // 2) Mint the secret if it isn't set yet — transparent, idempotent.
+      let secretSet = !!metaData.hmac_secret_set;
+      if (!secretSet) {
+        const bootRes = await fetch('/api/ai/connect-info/bootstrap', {
+          method: 'POST',
+        });
+        const bootData = await bootRes.json().catch(() => ({}));
+        if (!bootRes.ok || bootData?.ok === false) {
+          throw new Error(bootData?.detail || 'Could not generate agent credentials.');
+        }
+        secretSet = !!bootData.hmac_secret_set;
+      }
+
+      // 3) If the caller asked to see the secret, fetch it explicitly.
+      //    Otherwise the masked fingerprint is enough for the UI.
+      if (reveal && secretSet) {
+        const revealRes = await fetch('/api/ai/connect-info/reveal', {
+          method: 'POST',
+        });
+        const revealData = await revealRes.json().catch(() => ({}));
+        if (!revealRes.ok || revealData?.ok === false) {
+          throw new Error(revealData?.detail || 'Could not reveal agent credentials.');
+        }
+        setAgentSecret(revealData.hmac_secret || '');
+      } else {
+        setAgentSecret(metaData.masked_hmac_secret || '');
       }
-      setAgentSecret(data.hmac_secret || '');
-      setAgentTier(data.access_tier === 'full' ? 'full' : 'restricted');
       setAgentMsg({ type: 'ok', text: 'Agent key is ready. Copy it into your local or remote agent runtime.' });
     } catch (error) {
       setAgentMsg({

frontend/src/hooks/useAisUpstreamHealth.ts

@@ -0,0 +1,85 @@
+/**
+ * useAisUpstreamHealth — polls /api/health and exposes AIS proxy connectivity.
+ *
+ * Background: AISStream's WebSocket server went fully offline 2026-05-23 (TCP
+ * timeouts at stream.aisstream.io). The backend kept reconnecting in a tight
+ * loop and the ships layer silently went empty. Users had no signal that the
+ * problem was upstream, not their config. This hook surfaces the state so a
+ * banner can explain "AIS upstream is offline" instead of letting users
+ * wonder.
+ *
+ * The poll interval is intentionally relaxed (30s) — this is a low-urgency UX
+ * signal, not a real-time data feed. Backend already escalates top_status to
+ * "degraded" when AIS is configured-but-disconnected.
+ */
+import { useEffect, useRef, useState } from 'react';
+import { API_BASE } from '@/lib/api';
+
+export interface AisUpstreamHealth {
+  /** True when we've received a vessel message in the last ~60s. */
+  connected: boolean;
+  /** Seconds since the last vessel message; null when we've never seen one. */
+  lastMsgAgeSeconds: number | null;
+  /**
+   * True when the SPKI-pinned fallback is in effect (issue #258).
+   * Data still flows in this mode — it's a separate, less urgent signal
+   * than ``connected``.
+   */
+  degradedTls: boolean;
+  /** How many times the proxy has been spawned (sustained growth without
+   *  ``connected`` means upstream is dead and we're respawning in a loop). */
+  proxySpawnCount: number;
+  /** Whether the operator has configured an API key. When false, the banner
+   *  shouldn't fire because "AIS is off" is the intended state. The backend
+   *  signals this via the ``connected`` flag being false AND no msg ever
+   *  seen — we approximate it by requiring at least one spawn before
+   *  declaring an outage. */
+  aisEnabled: boolean;
+}
+
+const POLL_INTERVAL_MS = 30_000;
+
+export function useAisUpstreamHealth(): AisUpstreamHealth | null {
+  const [health, setHealth] = useState<AisUpstreamHealth | null>(null);
+  const cancelledRef = useRef(false);
+
+  useEffect(() => {
+    cancelledRef.current = false;
+
+    const fetchHealth = async () => {
+      try {
+        const res = await fetch(`${API_BASE}/api/health`, { cache: 'no-store' });
+        if (!res.ok) return;
+        const body = await res.json();
+        if (cancelledRef.current) return;
+        const proxy = body?.ais_proxy ?? {};
+        // ``proxy_spawn_count > 0`` is the cheapest "AIS is enabled" check:
+        // if the backend never spawned the proxy (no API key, opt-out env)
+        // we shouldn't ever show the outage banner. Once the proxy has
+        // spawned at least once we know the operator wants AIS data.
+        const spawns = Number(proxy.proxy_spawn_count ?? 0);
+        setHealth({
+          connected: Boolean(proxy.connected),
+          lastMsgAgeSeconds:
+            proxy.last_msg_age_seconds == null
+              ? null
+              : Number(proxy.last_msg_age_seconds),
+          degradedTls: Boolean(proxy.degraded_tls),
+          proxySpawnCount: spawns,
+          aisEnabled: spawns > 0,
+        });
+      } catch {
+        // Backend unreachable — separate problem. Banner not relevant.
+      }
+    };
+
+    void fetchHealth();
+    const interval = setInterval(() => void fetchHealth(), POLL_INTERVAL_MS);
+    return () => {
+      cancelledRef.current = true;
+      clearInterval(interval);
+    };
+  }, []);
+
+  return health;
+}

start-backend.js

@@ -76,6 +76,13 @@ function canRun(command, args) {
   return !result.error && result.status === 0;
 }
 
+function canRunBackendPython(pythonBin) {
+  return (
+    canRun(pythonBin, ["-V"]) &&
+    canRun(pythonBin, ["-c", "import fastapi, uvicorn"])
+  );
+}
+
 function findBasePython() {
   const candidates = isWindows
     ? [
@@ -135,12 +142,12 @@ function rebuildBackendVenv(targetDir, basePython) {
   if (result.error || result.status !== 0) {
     return null;
   }
-  return canRun(repairedBin, ["-V"]) ? repairedBin : null;
+  return canRunBackendPython(repairedBin) ? repairedBin : null;
 }
 
 function ensureBackendVenv() {
   for (const candidate of venvCandidates) {
-    if (fs.existsSync(candidate) && canRun(candidate, ["-V"])) {
+    if (fs.existsSync(candidate) && canRunBackendPython(candidate)) {
       persistSelectedVenv(candidate);
       return candidate;
     }