gstack/bin/gstack-brain-sync

#!/usr/bin/env bash
# gstack-brain-sync — drain queue, commit allowlisted paths, push to remote.
#
# Usage:
#   gstack-brain-sync --once           drain queue, commit, push (default)
#   gstack-brain-sync --status         print sync health as JSON
#   gstack-brain-sync --skip-file <p>  add <p> to ~/.gstack/.brain-skip.txt
#   gstack-brain-sync --drop-queue --yes   clear queue without committing
#   gstack-brain-sync --discover-new   scan allowlist dirs, enqueue changed files
#
# Invoked by the preamble at skill START and END boundaries. No persistent
# daemon. Typical run <1s when queue empty; ~200-800ms with network push.
#
# Singleton enforcement: flock on ~/.gstack/.brain-sync.lock. Concurrent
# invocations queue and serialize.
#
# Env:
#   GSTACK_HOME — override ~/.gstack (aligns with writers).

# Heredoc delivery guard. bash 5.2+ writes a heredoc body <=64KiB through a
# pipe in the forked child before exec, with no reader on the other end. On
# macOS under pipe-KVA pressure a fresh pipe gets a 512-byte buffer, so any
# body >=512B blocks write() forever and the script hangs at startup with no
# output. Compat level 50 restores the tempfile path. These scripts are
# bash-3.2-clean, so the compat level costs them nothing. Not exported: the
# guard is per-script, and it survives `bash script.sh` call sites that
# bypass the shebang.
BASH_COMPAT=50

set -uo pipefail

. "$(dirname "$0")/gstack-state-root.sh" 2>/dev/null || { echo "$0: cannot resolve the gstack state root: $(dirname "$0")/gstack-state-root.sh is missing. fix: reinstall with ./setup or /gstack-upgrade (docs/state-root.md)" >&2; exit 1; }
gstack_state_root_select
GSTACK_HOME="$_gstack_sr_root"
# Maildir-style spool: one FILE per record, <epoch>-<pid>-<uniq>.json.
# Writers (gstack-brain-enqueue, --discover-new) create records via tmp-file
# + atomic rename; the drain deletes exactly the files it snapshotted. No
# shared inode between writer and drainer → no append/rewrite race.
QUEUE_DIR="$GSTACK_HOME/.brain-queue.d"
# Legacy single-file queue: kept ONLY for migration. Pre-spool writers
# appended lines here; migrate_legacy_queue converts them to spool files.
QUEUE="$GSTACK_HOME/.brain-queue.jsonl"
ALLOWLIST="$GSTACK_HOME/.brain-allowlist"
PRIVACY_MAP="$GSTACK_HOME/.brain-privacy-map.json"
SKIP_FILE="$GSTACK_HOME/.brain-skip.txt"
STATUS_FILE="$GSTACK_HOME/.brain-sync-status.json"
LAST_PUSH_FILE="$GSTACK_HOME/.brain-last-push"
LOCK_FILE="$GSTACK_HOME/.brain-sync.lock"
DISCOVER_CURSOR="$GSTACK_HOME/.brain-discover-cursor"

SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
CONFIG_BIN="$SCRIPT_DIR/gstack-config"

# Egress receipt helpers (_receipted_git): receipt-before-send, fail-closed.
. "$SCRIPT_DIR/gstack-egress-lib.sh"

# origin host for receipt records (github.com etc).
remote_host() {
  local url host
  url=$(git -C "$GSTACK_HOME" remote get-url origin 2>/dev/null || echo "")
  host="${url#*://}"; host="${host#*@}"; host="${host%%[/:]*}"
  echo "${host:-unknown}"
}

# Remote-specific hint for auth errors (branch on origin URL).
remote_auth_hint() {
  local url
  url=$(git -C "$GSTACK_HOME" remote get-url origin 2>/dev/null || echo "")
  case "$url" in
    *github.com*|*@github.*) echo "run: gh auth status (and gh auth refresh if needed)" ;;
    *gitlab*) echo "run: glab auth status" ;;
    *) echo "check 'git remote -v' and your credentials" ;;
  esac
}

write_status() {
  # args: status_code message [extra_json_blob]
  local code="$1"
  local msg="$2"
  local extra="${3:-{\}}"
  local ts
  ts=$(date -u +%Y-%m-%dT%H:%M:%SZ 2>/dev/null || echo "")
  python3 - "$STATUS_FILE" "$code" "$msg" "$ts" "$extra" <<'PYEOF' 2>/dev/null || true
import json, sys
path, code, msg, ts, extra = sys.argv[1:6]
try:
    extra_obj = json.loads(extra) if extra else {}
except Exception:
    extra_obj = {}
data = {"status": code, "message": msg, "ts": ts, **extra_obj}
with open(path, "w") as f:
    json.dump(data, f)
    f.write("\n")
PYEOF
}

# Read config; return 0 if sync active, 1 otherwise.
sync_active() {
  if [ ! -d "$GSTACK_HOME/.git" ]; then
    return 1
  fi
  local mode
  mode=$("$CONFIG_BIN" get artifacts_sync_mode 2>/dev/null || echo off)
  [ "$mode" = "off" ] && return 1
  return 0
}

# Secret regex families — stdin scan. Exits 0 clean, 1 if hit.
# Echoes the matching pattern family name on hit. Uses python3 -c (not
# heredoc) so sys.stdin stays available for the diff content.
secret_scan_stdin() {
  python3 -c "
import sys, re
patterns = [
    ('aws-access-key', re.compile(r'AKIA[0-9A-Z]{16}')),
    ('github-token', re.compile(r'\\b(gh[pousr]_[A-Za-z0-9]{20,}|github_pat_[A-Za-z0-9_]{20,})')),
    ('openai-key', re.compile(r'\\bsk-[A-Za-z0-9_-]{20,}')),
    ('pem-block', re.compile(r'-----BEGIN [A-Z ]{3,}-----')),
    ('jwt', re.compile(r'\\beyJ[A-Za-z0-9_-]{10,}\\.[A-Za-z0-9_-]{10,}\\.[A-Za-z0-9_-]{10,}\\b')),
    ('bearer-token-json',
     # JSON-embedded auth headers. The optional Bearer/Basic/Token prefix
     # matters: real auth values include a literal space after the scheme
     # name, but the value charset below does not include spaces, so
     # without the optional prefix every Bearer token in a JSON blob slips
     # past the scanner.
     re.compile(r'\"(authorization|api[_-]?key|apikey|token|secret|password)\"\\s*:\\s*\"(Bearer |Basic |Token )?[A-Za-z0-9_./+=-]{16,}\"',
                re.IGNORECASE)),
]
text = sys.stdin.read()
for name, rx in patterns:
    m = rx.search(text)
    if m:
        snippet = m.group(0)
        if len(snippet) > 30:
            snippet = snippet[:30] + '...'
        print(name + ':' + snippet)
        sys.exit(1)
sys.exit(0)
"
}

# True (0) if the spool holds at least one record file.
spool_has_records() {
  local f
  for f in "$QUEUE_DIR"/*.json; do
    [ -e "$f" ] && return 0
  done
  return 1
}

# Convert one legacy queue file's lines into spool record files (tmp +
# os.replace, one file per line). Reads the file TWICE before unlinking: a
# pre-rename writer can still append through its already-open fd after our
# rename, and those appends land in the renamed file — the second pass
# NARROWS the tail-race window (transition-only: it applies to pre-spool
# writers, and a writer that appends after the second read but before the
# unlink can still lose that line; spool-native writers are immune).
# Unparseable lines migrate as-is; finalize_queue quarantines + warns on them.
convert_legacy_file() {
  local legacy="$1"
  python3 - "$legacy" "$QUEUE_DIR" <<'PYEOF' 2>/dev/null || true
import os, sys, time

legacy, spool = sys.argv[1:3]

def read_lines(path):
    try:
        with open(path) as f:
            return [l.rstrip("\r\n") for l in f if l.strip()]
    except (FileNotFoundError, OSError):
        return []

seq = 0
def write_spool(line):
    global seq
    seq += 1
    tmp = os.path.join(spool, f".tmp-{os.getpid()}-m{seq}")
    with open(tmp, "w") as f:
        f.write(line + "\n")
    os.replace(tmp, os.path.join(spool, f"{int(time.time())}-{os.getpid()}-m{seq}.json"))

written = set()
for _pass in (1, 2):   # second read closes the pre-rename-fd tail race
    for line in read_lines(legacy):
        if line not in written:   # identical duplicates collapse, as the old rewrite did
            write_spool(line)
            written.add(line)
os.unlink(legacy)
PYEOF
}

# Legacy migration (transition window only). If the single-file queue holds
# records, atomically rename it aside and convert each line to a spool file.
# A concurrent OLD writer that recreates a fresh legacy file after the rename
# simply gets migrated on the NEXT drain — nothing is lost, only deferred one
# boundary. Runs inside the run lock, before the drain reads the spool.
migrate_legacy_queue() {
  local migrating="$QUEUE.migrating"
  # Crash leftover: a prior migration renamed but died before unlink. Some of
  # its lines may already exist as spool files — re-converting duplicates is
  # safe (at-least-once; the drain dedups paths per snapshot and downstream
  # content-hash dedup absorbs re-syncs). Losing the file would not be. If
  # the conversion itself fails, the file stays for the next run (never rm a
  # non-empty .migrating file outside convert_legacy_file's own unlink).
  if [ -f "$migrating" ]; then
    if [ -s "$migrating" ]; then
      mkdir -p "$QUEUE_DIR" 2>/dev/null || return 0
      convert_legacy_file "$migrating"
    else
      rm -f "$migrating" 2>/dev/null || true
    fi
  fi
  # If the leftover STILL holds records, the conversion failed (e.g. python3
  # unavailable). The mv below would overwrite it and destroy those records —
  # exactly the never-destroy invariant above. Defer this run's migration;
  # the next run retries both files.
  [ -s "$migrating" ] && return 0
  if [ -s "$QUEUE" ]; then
    mkdir -p "$QUEUE_DIR" 2>/dev/null || return 0
    mv -f "$QUEUE" "$migrating" 2>/dev/null || return 0
    convert_legacy_file "$migrating"
  fi
  return 0
}

# Compute matched allowlisted, privacy-filtered path set from the spool.
# Output: newline-delimited relative paths that should be staged.
#
# #2549: every non-staged queue entry is CLASSIFIED, never silently discarded.
# When $2 is given, a JSON classification lands there:
#   {"retained": [privacy/mode-held paths that stay queued],
#    "dropped": {"skipped": [...], "invalid": [...], "unmatched": [...], "missing": [...]}}
# retained entries would sync if the user raises artifacts_sync_mode, so they
# stay in the queue; dropped classes can never sync (explicit skip, escape
# attempt, no allowlist glob, not on disk) and are removed WITH a counted
# status — the old behavior truncated the whole queue and reported every one
# of these, including privacy holds, as "no allowlisted changes".
#
# Spool snapshot ($3): the sorted list of spool record filenames read here is
# written to the snapshot manifest, one filename per line. finalize_queue
# deletes exactly the manifest's files and never touches records created
# after this listing — a concurrent enqueue is a separate file by
# construction, so it simply rides to the next drain.
compute_paths_to_stage() {
  local mode="$1"
  local class_file="${2:-}"
  local snapshot_file="${3:-}"
  python3 - "$GSTACK_HOME" "$QUEUE_DIR" "$ALLOWLIST" "$PRIVACY_MAP" "$SKIP_FILE" "$mode" "$class_file" "$snapshot_file" <<'PYEOF'
import sys, json, os, fnmatch, glob

gstack_home, spool_dir, allowlist_path, privacy_path, skip_path, mode, class_file, snapshot_file = sys.argv[1:9]

def load_lines(path):
    try:
        with open(path) as f:
            return [l.strip() for l in f if l.strip() and not l.lstrip().startswith("#")]
    except FileNotFoundError:
        return []

def load_privacy_map(path):
    # Returns (entries, corrupt). Non-dict entries are filtered out
    # defensively — the map may be PULLED from the artifacts remote, so a
    # malformed entry like ["bad"] is remotely triggerable and used to raise
    # mid-classification (after the snapshot manifest was written), which the
    # old finalize turned into a full queue wipe. Any malformed shape also
    # marks the map CORRUPT: privacy classification cannot be trusted, so the
    # caller holds every queued record instead of guessing (a corrupt privacy
    # map silently treated as empty would over-share behavioral data).
    try:
        with open(path) as f:
            data = json.load(f)
    except FileNotFoundError:
        return [], False
    except json.JSONDecodeError:
        return [], True
    if not isinstance(data, list):
        return [], True
    # Expected: [{"pattern": "glob", "class": "artifact" | "behavioral"}]
    entries = [e for e in data if isinstance(e, dict)]
    return entries, len(entries) != len(data)

allowlist_globs = load_lines(allowlist_path)
privacy_map, privacy_corrupt = load_privacy_map(privacy_path)
# Normalize skip entries to the POSIX form queued paths use, so a backslash
# entry in .brain-skip.txt still matches on Windows. The drain is the safety
# boundary that actually stages files, so it must normalize identically to
# discover_new — otherwise an explicitly-skipped file gets committed.
skip_lines = {s.replace(os.sep, "/") for s in load_lines(skip_path)}

# Snapshot the spool: sorted (= chronological, filenames are epoch-first)
# list of record files at read time. Records that appear after this listing
# belong to the NEXT drain. Files we cannot read stay OUT of the manifest so
# finalize never deletes a record this drain didn't actually consume.
try:
    snapshot = sorted(n for n in os.listdir(spool_dir) if n.endswith(".json"))
except (FileNotFoundError, NotADirectoryError):
    snapshot = []

queue_paths = set()
consumed = []
for name in snapshot:
    try:
        with open(os.path.join(spool_dir, name)) as f:
            line = f.readline().strip()
    except OSError:
        continue
    consumed.append(name)
    if not line:
        continue
    try:
        obj = json.loads(line)
        p = obj.get("file")
        if isinstance(p, str):
            queue_paths.add(p)
    except json.JSONDecodeError:
        continue   # unparseable record: finalize keeps + warns

if snapshot_file:
    with open(snapshot_file, "w") as f:
        for name in consumed:
            f.write(name + "\n")

def path_matches_any(path, globs):
    for pattern in globs:
        if fnmatch.fnmatchcase(path, pattern):
            return True
    return False

def privacy_class(path, mapping):
    for entry in mapping:
        pat = entry.get("pattern")
        if pat and fnmatch.fnmatchcase(path, pat):
            return entry.get("class", "artifact")
    # Default class when no pattern matches: artifact (safe default).
    return "artifact"

# mode filter: 'off' → nothing; 'artifacts-only' → only artifact class;
# 'full' → both classes.
def mode_allows(cls, mode):
    if mode == "off":
        return False
    if mode == "artifacts-only":
        return cls == "artifact"
    return True  # full

final = []
classified = {"retained": [], "dropped": {"skipped": [], "invalid": [], "unmatched": [], "missing": []}}
if privacy_corrupt:
    # Fail-safe: with an untrustworthy privacy map, stage NOTHING and drop
    # NOTHING — retain every queued record until the map is fixed. The next
    # drain re-classifies from scratch.
    print("BRAIN_SYNC: warning: privacy map at " + privacy_path +
          " is malformed — holding all queued records until it is fixed", file=sys.stderr)
    classified["retained"] = sorted(queue_paths)
    queue_paths = set()
for p in sorted(queue_paths):
    if p in skip_lines:
        classified["dropped"]["skipped"].append(p)
        continue
    # Must be under GSTACK_HOME root. Reject absolute + reject ../ escape.
    if p.startswith("/") or ".." in p.split("/"):
        classified["dropped"]["invalid"].append(p)
        continue
    # Must match at least one allowlist glob.
    if not path_matches_any(p, allowlist_globs):
        classified["dropped"]["unmatched"].append(p)
        continue
    # Must survive privacy mode filter — held entries STAY QUEUED (retained):
    # they would sync under a higher artifacts_sync_mode, and reporting them
    # as "no allowlisted changes" was #2549's misattribution.
    cls = privacy_class(p, privacy_map)
    if not mode_allows(cls, mode):
        classified["retained"].append(p)
        continue
    # Must exist on disk — can't stage what isn't there.
    if not os.path.exists(os.path.join(gstack_home, p)):
        classified["dropped"]["missing"].append(p)
        continue
    final.append(p)

if class_file:
    with open(class_file, "w") as f:
        json.dump(classified, f)

for p in final:
    print(p)
PYEOF
}

# Finalize the drain: delete exactly the spool record files this drain
# consumed (per the snapshot manifest) AND positively classified. Deletion is
# EXPLICIT-DELETE-ONLY: a record is unlinked only when its path appears in
# (staged paths ∪ classified dropped). The old polarity ("delete unless
# retained") turned a missing/unparseable classification into retained=∅ and
# wiped every snapshotted record — remotely triggerable via a malformed
# pulled privacy map that raised AFTER the manifest write. Now a
# missing/unparseable class_file or paths_file deletes NOTHING (warn +
# return), and a path the classification never mentions stays queued.
# The predecessor (a shared-file queue rewrite) had a lockless-append race
# between its live re-read and the os.replace; with one file per record that
# race class is structurally gone — a concurrent enqueue is a separate file
# the snapshot never listed, so finalize cannot touch it. Crash semantics are
# at-least-once: a drain that dies before finalize leaves its spool files in
# place and the next run re-drains them; downstream content-hash dedup
# absorbs the duplicates. Unparseable records move to $QUEUE_DIR/quarantine/
# (never deleted) so they stop re-warning at every boundary. Dropped-path
# detail goes to a 0600 sidecar so the status line can stay content-free
# (counts only).
finalize_queue() {
  local snapshot_file="$1"  # spool filenames this drain consumed, one per line
  local class_file="$2"     # classification JSON from compute_paths_to_stage
  local paths_file="$3"     # staged paths (compute_paths_to_stage stdout), one per line
  # Fail-open by design (a failed finalize self-corrects next run: re-stage →
  # nothing-to-commit), but say so — a silent failure here would let the
  # subsequent "ok/idle" status claim a drain that did not happen.
  python3 - "$QUEUE_DIR" "$snapshot_file" "$class_file" "$paths_file" "$GSTACK_HOME/.brain-sync-drops.json" <<'PYEOF' || echo "BRAIN_SYNC: warning: queue finalize failed — entries retained; next run re-drains" >&2
import json, os, sys, time
spool_dir, snapshot_file, class_file, paths_file, drops_file = sys.argv[1:6]

def lines(path):
    try:
        with open(path) as f:
            return [l.rstrip("\r\n") for l in f if l.strip()]
    except FileNotFoundError:
        return []

# Explicit-delete-only inputs. Either input unreadable → delete NOTHING.
try:
    with open(class_file) as f:
        classified = json.load(f)
    if not isinstance(classified, dict):
        raise ValueError("classification is not an object")
except Exception:
    print("BRAIN_SYNC: warning: classification unreadable — no queue records deleted; next run re-drains", file=sys.stderr)
    sys.exit(0)
try:
    with open(paths_file) as f:
        staged = {l.strip() for l in f if l.strip()}
except Exception:
    print("BRAIN_SYNC: warning: staged-paths file unreadable — no queue records deleted; next run re-drains", file=sys.stderr)
    sys.exit(0)

dropped = set()
for group in (classified.get("dropped", {}) or {}).values():
    dropped.update(group)
deletable = staged | dropped

unparseable = 0
for name in lines(snapshot_file):
    full = os.path.join(spool_dir, name)
    try:
        with open(full) as f:
            rec = f.readline().strip()
    except OSError:
        continue                 # unreadable now: leave it for the next drain
    p = None
    try:
        p = json.loads(rec).get("file")
    except Exception:
        pass
    if not isinstance(p, str):
        # Never destroy what we can't read — but don't leave it re-warning at
        # every boundary either: move it aside for inspection.
        unparseable += 1
        try:
            qdir = os.path.join(spool_dir, "quarantine")
            os.makedirs(qdir, exist_ok=True)
            os.replace(full, os.path.join(qdir, name))
        except OSError:
            pass                 # quarantine move failed — leave in place; next run retries
        continue
    if p not in deletable:
        continue                 # retained / unclassified: stays queued (explicit-delete-only)
    try:
        os.unlink(full)          # staged or dropped: fully processed
    except FileNotFoundError:
        pass
if unparseable:
    print(f"BRAIN_SYNC: {unparseable} unparseable spool record(s) moved to quarantine (inspect {os.path.join(spool_dir, 'quarantine')})", file=sys.stderr)

if dropped:
    fd = os.open(drops_file, os.O_WRONLY | os.O_CREAT | os.O_TRUNC, 0o600)
    with os.fdopen(fd, "w") as f:
        json.dump({"ts": time.strftime("%Y-%m-%dT%H:%M:%SZ", time.gmtime()),
                   "dropped": classified.get("dropped", {})}, f)
PYEOF
}

# Human-readable classification counts for status messages.
queue_summary() {
  local class_file="$1"
  python3 - "$class_file" <<'PYEOF' 2>/dev/null || echo ""
import json, sys
try:
    with open(sys.argv[1]) as f:
        c = json.load(f)
except Exception:
    print(""); sys.exit(0)
d = c.get("dropped", {}) or {}
parts = []
r = len(c.get("retained", []))
if r: parts.append(f"{r} privacy-held retained")
for k in ("skipped", "unmatched", "missing", "invalid"):
    n = len(d.get(k, []))
    if n: parts.append(f"{n} {k} dropped")
print("; ".join(parts))
PYEOF
}

subcmd_once() {
  if ! sync_active; then
    # Silent no-op when feature not initialized / disabled.
    exit 0
  fi

  # Singleton lock via atomic mkdir. `flock(1)` isn't on macOS by default;
  # `mkdir` is atomic on every POSIX filesystem. If another --once is already
  # running, skip (don't wait) — the next skill boundary will catch up.
  local lock_dir="${LOCK_FILE}.d"
  if ! mkdir "$lock_dir" 2>/dev/null; then
    # Is the lock stale? Check the pidfile inside. If process is dead, clear it.
    if [ -f "$lock_dir/pid" ]; then
      local lock_pid
      lock_pid=$(cat "$lock_dir/pid" 2>/dev/null || echo "")
      if [ -n "$lock_pid" ] && ! kill -0 "$lock_pid" 2>/dev/null; then
        # Stale lock — clear and retry once.
        rm -rf "$lock_dir" 2>/dev/null || true
        if ! mkdir "$lock_dir" 2>/dev/null; then
          exit 0
        fi
      else
        # Lock is held by a live process.
        exit 0
      fi
    else
      # Lock dir without pidfile — treat as held; don't touch.
      exit 0
    fi
  fi
  echo "$$" > "$lock_dir/pid" 2>/dev/null || true
  # Release the lock on EVERY exit from here on — including the empty-queue
  # fast path and an INT during the detector's network push. Leaking it would
  # rely on next-run stale-pid detection, which PID reuse can defeat (kill -0
  # matching an unrelated live process wedges sync at every boundary). The
  # mktemp block below re-traps with tempfile cleanup added; both traps keep
  # the lock removal.
  trap 'rm -rf "$lock_dir" 2>/dev/null || true' EXIT INT TERM

  # Convert any legacy single-file queue lines into spool records before the
  # drain reads the spool (transition window for pre-spool writers).
  migrate_legacy_queue

  # Janitor: reap orphaned enqueue temp files. A writer killed between its
  # tmp write and the atomic rename leaves `.tmp-*` behind forever — it never
  # becomes a record and nothing else touches it. One hour is far beyond any
  # live writer's write→rename window, so a fresh tmp (an in-flight enqueue)
  # is never touched. Runs inside the run lock, so it can't race the drain.
  find "$QUEUE_DIR" -maxdepth 1 -type f -name '.tmp-*' -mmin +60 -delete 2>/dev/null || true

  local mode
  mode=$("$CONFIG_BIN" get artifacts_sync_mode 2>/dev/null || echo off)

  # #2516: advance the brain worktree gbrain indexes to the artifacts repo's
  # HEAD once a day — previously it only moved when setup-gbrain / sync-gbrain
  # / brain-restore ran, so brains silently served stale code forever. Runs
  # inside THIS run lock (never concurrent with the ingest steps below) and
  # before they touch the worktree. Attempt-throttled: the stamp is written on
  # ATTEMPT, so a persistently-failing advance warns once per 24h, not at
  # every skill boundary. The advance itself refuses dirty or unmanaged
  # worktrees and never force-removes (see gstack-gbrain-source-wireup).
  if [ -e "${GSTACK_BRAIN_WORKTREE:-$HOME/.gstack-brain-worktree}" ]; then
    local adv_stamp adv_now adv_last adv_age
    adv_stamp="$GSTACK_HOME/.brain-worktree-last-advance"
    adv_now=$(date +%s)
    adv_last=$(cat "$adv_stamp" 2>/dev/null || echo 0)
    case "$adv_last" in ''|*[!0-9]*) adv_last=0 ;; esac
    adv_age=$(( adv_now - adv_last ))
    if [ "$adv_age" -ge 86400 ]; then
      echo "$adv_now" > "$adv_stamp" 2>/dev/null || true
      if ! "$SCRIPT_DIR/gstack-gbrain-source-wireup" --advance-only 1>&2; then
        echo "BRAIN_SYNC: warning: brain worktree advance failed — gbrain may be indexing stale code (run gstack-gbrain-source-wireup to repair)" >&2
      fi
    fi
  fi

  # #2549 unpushed-commit detector: a prior drain may have COMMITTED but
  # failed to push (auth blip, offline). The data was never lost — it sits in
  # a local commit — but nothing re-pushed it until NEW changes arrived.
  # Retry the push up front, inside the lock. Receipted fail-closed like
  # every other push; a receipt REFUSAL skips the retry without blocking the
  # rest of the drain (local staging must not wedge on receipt problems).
  # Guards: origin/<branch> may not exist yet (first sync, deleted remote).
  #
  # Throttled: the preamble runs --once at EVERY skill boundary, so an
  # unthrottled retry would pay a full network push attempt per boundary in
  # exactly the steady states this targets (offline, broken auth) — and a
  # captive-portal push can block 30-75s against the header's "<1s when
  # idle" promise. Attempts are recorded (success or fail) and retried at
  # most every 10 minutes; the push itself never prompts for credentials and
  # bounds stalled transfers via git's own low-speed limits (portable — stock
  # macOS ships no `timeout` binary).
  #
  # Author-scoped — EXCLUSIVELY: `git push origin HEAD` publishes every
  # unpushed commit, so the retry fires only when ALL unpushed commits are
  # gstack-brain-sync's own. One interleaved user commit disables the
  # auto-retry entirely (adversarial review: an existential check would
  # silently auto-publish a user's manual ~/.gstack commit the moment a bot
  # commit sat in front of it). User commits ride along when a REAL drain
  # pushes, as before — the detector never publishes work it didn't create.
  local det_branch det_unpushed det_total det_now det_last
  det_branch=$(git -C "$GSTACK_HOME" rev-parse --abbrev-ref HEAD 2>/dev/null || echo "")
  # Detached HEAD reads as the literal "HEAD" — origin/HEAD usually resolves,
  # so without this exclusion the detector would retry a doomed push forever.
  [ "$det_branch" = "HEAD" ] && det_branch=""
  if [ -n "$det_branch" ] && git -C "$GSTACK_HOME" rev-parse --verify --quiet "origin/$det_branch" >/dev/null 2>&1; then
    det_unpushed=$(git -C "$GSTACK_HOME" rev-list --count --author="gstack-brain-sync" "origin/$det_branch..HEAD" 2>/dev/null || echo 0)
    det_total=$(git -C "$GSTACK_HOME" rev-list --count "origin/$det_branch..HEAD" 2>/dev/null || echo 0)
    case "$det_unpushed" in ''|*[!0-9]*) det_unpushed=0 ;; esac
    case "$det_total" in ''|*[!0-9]*) det_total=0 ;; esac
    det_now=$(date +%s)
    det_last=$(cat "$GSTACK_HOME/.brain-last-push-attempt" 2>/dev/null || echo 0)
    case "$det_last" in ''|*[!0-9]*) det_last=0 ;; esac
    if [ "$det_unpushed" -gt 0 ] && [ "$det_unpushed" -eq "$det_total" ] && [ $(( det_now - det_last )) -ge 600 ]; then
      echo "$det_now" > "$GSTACK_HOME/.brain-last-push-attempt" 2>/dev/null || true
      local det_host
      det_host=$(remote_host)
      if GSTACK_HOME="$GSTACK_HOME" _receipted_git closed brain-sync "$det_host" curated-memory-git-push "artifacts_sync_mode!=off" \
          bash -c 'GIT_TERMINAL_PROMPT=0 git -c http.lowSpeedLimit=1024 -c http.lowSpeedTime=30 -C "$1" push origin HEAD 2>/dev/null' _ "$GSTACK_HOME"; then
        date -u +%Y-%m-%dT%H:%M:%SZ > "$LAST_PUSH_FILE"
      fi
    fi
  fi

  # Empty-queue fast path: this is the steady state at every skill boundary.
  # Skipping compute/finalize here is safe — with zero spool records there is
  # nothing to classify, retain, or drop, and a record created after this
  # check simply waits for the next boundary. The legacy file is checked too:
  # an OLD writer may have recreated it after the migration above (it gets
  # migrated next run, but the depth is honest now) — and so is a leftover
  # .migrating file: if its conversion failed above (e.g. python3 missing),
  # records are still pending, so "idle" would be dishonest. (The detector
  # above already ran: its whole point is re-pushing stranded commits when
  # the queue is empty.) The lock-release trap installed at acquisition
  # covers this exit.
  if ! spool_has_records && [ ! -s "$QUEUE" ] && [ ! -s "$QUEUE.migrating" ]; then
    write_status "idle" "queue empty"
    exit 0
  fi

  local paths_file class_file snapshot_file
  paths_file=$(mktemp /tmp/brain-sync-paths.XXXXXX) || { rm -rf "$lock_dir" 2>/dev/null; write_status "error" "mktemp failed"; exit 1; }
  class_file=$(mktemp /tmp/brain-sync-class.XXXXXX) || { rm -f "$paths_file"; rm -rf "$lock_dir" 2>/dev/null; write_status "error" "mktemp failed"; exit 1; }
  snapshot_file=$(mktemp /tmp/brain-sync-snapshot.XXXXXX) || { rm -f "$paths_file" "$class_file"; rm -rf "$lock_dir" 2>/dev/null; write_status "error" "mktemp failed"; exit 1; }
  # Single trap covers all: lock cleanup AND tempfile cleanup.
  trap 'rm -f "$paths_file" "$class_file" "$snapshot_file" 2>/dev/null; rm -rf "$lock_dir" 2>/dev/null || true' EXIT INT TERM

  # Fail-safe (G1): a classifier that dies mid-run (ENOSPC/OOM/SIGKILL, or a
  # shape the defensive filters don't cover) may have already written the
  # snapshot manifest but no classification. Finalizing on that state is what
  # used to wipe the queue — so on a nonzero exit, warn loudly, do NOT call
  # finalize_queue, and leave everything queued for the next drain.
  if ! compute_paths_to_stage "$mode" "$class_file" "$snapshot_file" > "$paths_file"; then
    echo "BRAIN_SYNC: warning: queue classification failed — no records consumed; next run re-drains" >&2
    write_status "error" "classification failed; queue preserved (next run retries)"
    exit 0
  fi
  if [ ! -s "$paths_file" ]; then
    # Nothing stageable. Finalize the snapshot (retained entries survive;
    # classified drops removed; records created after the snapshot untouched).
    finalize_queue "$snapshot_file" "$class_file" "$paths_file"
    local summary
    summary=$(queue_summary "$class_file")
    write_status "idle" "no stageable changes${summary:+ ($summary)}"
    exit 0
  fi

  # Stage with git add -f (forces past .gitignore=*) explicit paths only.
  while IFS= read -r p; do
    p="${p%$'\r'}"   # Windows: compute_paths_to_stage's python print() emits CRLF;
                     # a trailing CR makes the pathspec match nothing (silent no-stage).
    [ -z "$p" ] && continue
    git -C "$GSTACK_HOME" add -f -- "$p" 2>/dev/null || true
  done < "$paths_file"

  # Secret-scan staged diff.
  local scan_out
  scan_out=$(git -C "$GSTACK_HOME" diff --cached 2>/dev/null | secret_scan_stdin || true)
  if [ -n "$scan_out" ]; then
    # Hit — unstage, preserve queue, write loud status.
    git -C "$GSTACK_HOME" reset HEAD -- . >/dev/null 2>&1 || true
    local hint
    hint="secret pattern detected ($scan_out). Remediation: review the staged file, then run: gstack-brain-sync --skip-file <path> OR edit the content."
    write_status "blocked" "$hint"
    echo "BRAIN_SYNC: blocked: $scan_out" >&2
    exit 0
  fi

  # Egress receipt for the push, written BEFORE the commit consumes the
  # queue (amendment C7 ordering): a refused receipt exits HERE, before any
  # queue mutation or local commit, so the queue stays intact and the next
  # run retries the whole drain. Content-free: git owns the bytes
  # (sha256:null). Fail-closed.
  local push_host receipt_err
  push_host=$(remote_host)
  if ! receipt_err=$(GSTACK_HOME="$GSTACK_HOME" "$SCRIPT_DIR/gstack-egress-receipt" write \
      --sink brain-sync --host "$push_host" --class curated-memory-git-push \
      --no-payload --consent "artifacts_sync_mode!=off" 2>&1 >/dev/null); then
    write_status "push_failed" "EGRESS_RECEIPT_FAILED: receipt not writable; push refused (queue preserved)"
    _gstack_egress_refusal "brain-sync push" "$(printf '%s' "$receipt_err" | head -c 300)"
    exit 1
  fi

  # Commit with template message.
  local n ts
  n=$(wc -l < "$paths_file" | tr -d ' ')
  ts=$(date -u +%Y-%m-%dT%H:%M:%SZ)
  local msg="sync: $n file(s) | $ts"
  git -C "$GSTACK_HOME" -c user.email="gstack@localhost" -c user.name="gstack-brain-sync" \
    commit -q -m "$msg" 2>/dev/null || {
      # Nothing to commit (e.g. all files already committed). The drained
      # records leave the spool; retained + post-snapshot records survive.
      finalize_queue "$snapshot_file" "$class_file" "$paths_file"
      write_status "idle" "queue drained but no new changes to commit"
      exit 0
    }

  # Push. On reject, fetch + merge (merge driver handles JSONL) + retry once.
  local push_err
  push_err=$(git -C "$GSTACK_HOME" push origin HEAD 2>&1 >/dev/null) || {
    # Check if this is an auth error first — no point retrying.
    if echo "$push_err" | grep -qiE "auth|permission|403|401|forbidden"; then
      local hint
      hint=$(remote_auth_hint)
      write_status "push_failed" "push failed: auth error; commit retained locally, will retry next run. fix: $hint"
      echo "BRAIN_SYNC: push failed: auth. fix: $hint" >&2
      # Drained records leave the spool — they live in the local commit, which
      # the run-start detector re-pushes next time (#2549). Retained +
      # post-snapshot records survive the finalize.
      finalize_queue "$snapshot_file" "$class_file" "$paths_file"
      exit 0
    fi

    # Try a fetch-and-merge + retry. The fetch and the retry push are their
    # own attempted-egress ops, each receipted fail-closed (a refusal falls
    # through to the push_failed path below).
    if GSTACK_HOME="$GSTACK_HOME" _receipted_git closed brain-sync "$push_host" curated-memory-git-fetch "artifacts_sync_mode!=off" \
        bash -c 'git -C "$1" fetch origin 2>/dev/null' _ "$GSTACK_HOME"; then
      local branch
      branch=$(git -C "$GSTACK_HOME" rev-parse --abbrev-ref HEAD 2>/dev/null || echo main)
      if git -C "$GSTACK_HOME" merge --no-edit "origin/$branch" >/dev/null 2>&1; then
        if GSTACK_HOME="$GSTACK_HOME" _receipted_git closed brain-sync "$push_host" curated-memory-git-push "artifacts_sync_mode!=off" \
            bash -c 'git -C "$1" push origin HEAD 2>/dev/null' _ "$GSTACK_HOME"; then
          finalize_queue "$snapshot_file" "$class_file" "$paths_file"
          date -u +%Y-%m-%dT%H:%M:%SZ > "$LAST_PUSH_FILE"
          write_status "ok" "pushed $n file(s) after rebase"
          exit 0
        fi
      fi
    fi
    # Commit exists locally; the run-start detector re-pushes it next time.
    write_status "push_failed" "push failed: $(printf '%s' "$push_err" | head -1); commit retained locally, will retry next run"
    finalize_queue "$snapshot_file" "$class_file" "$paths_file"
    exit 0
  }

  # Success: drained records leave the spool (retained + post-snapshot survive).
  finalize_queue "$snapshot_file" "$class_file" "$paths_file"
  date -u +%Y-%m-%dT%H:%M:%SZ > "$LAST_PUSH_FILE"
  write_status "ok" "pushed $n file(s)"
  exit 0
}

subcmd_status() {
  if [ -f "$STATUS_FILE" ]; then
    cat "$STATUS_FILE"
  else
    echo '{"status":"unknown","message":"no status file yet"}'
  fi
  # Supplemental info (not in status file). Depth = spool record files plus
  # any not-yet-migrated legacy queue lines (transition window), including a
  # crash-leftover .migrating file — its records are still pending too.
  local queue_depth spool_depth legacy_depth
  spool_depth=$(ls "$QUEUE_DIR"/*.json 2>/dev/null | wc -l | tr -d ' ')
  legacy_depth=0
  [ -f "$QUEUE" ] && legacy_depth=$(wc -l < "$QUEUE" | tr -d ' ')
  [ -f "$QUEUE.migrating" ] && legacy_depth=$(( legacy_depth + $(wc -l < "$QUEUE.migrating" | tr -d ' ') ))
  queue_depth=$(( spool_depth + legacy_depth ))
  local last_push="never"
  [ -f "$LAST_PUSH_FILE" ] && last_push=$(cat "$LAST_PUSH_FILE" 2>/dev/null || echo never)
  local mode
  mode=$("$CONFIG_BIN" get artifacts_sync_mode 2>/dev/null || echo off)
  printf '{"queue_depth":%s,"last_push":"%s","mode":"%s"}\n' "$queue_depth" "$last_push" "$mode"
}

subcmd_skip_file() {
  local path="${1:-}"
  if [ -z "$path" ]; then
    echo "Usage: gstack-brain-sync --skip-file <path>" >&2
    exit 1
  fi
  mkdir -p "$GSTACK_HOME"
  # Avoid duplicate entries.
  if [ -f "$SKIP_FILE" ] && grep -Fxq "$path" "$SKIP_FILE"; then
    echo "already in skip list: $path"
    exit 0
  fi
  echo "$path" >> "$SKIP_FILE"
  echo "added to skip list: $path"
  echo "(future writers will not enqueue this path; existing queue entries ignored on next --once)"
}

subcmd_drop_queue() {
  local force="${1:-}"
  if [ "$force" != "--yes" ]; then
    echo "Refusing: --drop-queue discards pending syncs. Pass --yes to confirm." >&2
    exit 1
  fi
  # Remove spool record files, then truncate any legacy queue remnant —
  # including a crash-leftover .migrating file, whose records would otherwise
  # resurrect on the next drain via migrate_legacy_queue after the user
  # explicitly discarded the queue.
  local n=0 f
  for f in "$QUEUE_DIR"/*.json; do
    [ -e "$f" ] || continue
    rm -f "$f" 2>/dev/null && n=$(( n + 1 ))
  done
  if [ -f "$QUEUE" ]; then
    local legacy_n
    legacy_n=$(wc -l < "$QUEUE" | tr -d ' ')
    n=$(( n + legacy_n ))
    : > "$QUEUE"
  fi
  if [ -f "$QUEUE.migrating" ]; then
    local mig_n
    mig_n=$(wc -l < "$QUEUE.migrating" | tr -d ' ')
    n=$(( n + mig_n ))
    rm -f "$QUEUE.migrating" 2>/dev/null || true
  fi
  if [ "$n" -eq 0 ]; then
    echo "queue already empty"
    exit 0
  fi
  echo "dropped $n queue entries"
}

subcmd_discover_new() {
  if ! sync_active; then
    exit 0
  fi
  # Walk allowlist globs; enqueue any file where mtime+size differs from cursor.
  python3 - "$GSTACK_HOME" "$ALLOWLIST" "$DISCOVER_CURSOR" <<'PYEOF' 2>/dev/null || true
import sys, os, json, fnmatch, time
from datetime import datetime, timezone

gstack_home, allowlist_path, cursor_path = sys.argv[1:4]
spool_dir = os.path.join(gstack_home, ".brain-queue.d")
skip_path = os.path.join(gstack_home, ".brain-skip.txt")

def load_lines(path):
    try:
        with open(path) as f:
            return [l.strip() for l in f if l.strip() and not l.lstrip().startswith("#")]
    except FileNotFoundError:
        return []

def load_cursor(path):
    try:
        with open(path) as f:
            return json.load(f)
    except (FileNotFoundError, json.JSONDecodeError):
        return {}

def save_cursor(path, data):
    try:
        with open(path, "w") as f:
            json.dump(data, f)
    except OSError:
        pass

allowlist = load_lines(allowlist_path)
# Normalize skip entries to the same POSIX form as `rel` below, so a
# backslash entry in .brain-skip.txt still matches a normalized path on Windows.
skip = {s.replace(os.sep, "/") for s in load_lines(skip_path)}
cursor = load_cursor(cursor_path)
new_cursor = dict(cursor)
to_enqueue = []

# Walk all files under gstack_home, match against allowlist.
for root, dirs, files in os.walk(gstack_home):
    # Skip .git and .brain-* state files.
    if ".git" in root.split(os.sep):
        continue
    for name in files:
        full = os.path.join(root, name)
        # Repo paths are POSIX-relative. os.path.relpath yields backslash
        # separators on Windows, which never match the forward-slash allowlist
        # globs (e.g. "projects/*/learnings.jsonl"), so discovery silently
        # enqueued nothing under projects/ on Windows. Normalize to "/".
        rel = os.path.relpath(full, gstack_home).replace(os.sep, "/")
        if rel.startswith(".brain-"):
            continue
        if not any(fnmatch.fnmatchcase(rel, pat) for pat in allowlist):
            continue
        if rel in skip:
            continue
        try:
            st = os.stat(full)
            key = f"{int(st.st_mtime)}:{st.st_size}"
        except OSError:
            continue
        if cursor.get(rel) != key:
            to_enqueue.append((rel, key))

# Write spool records directly. The previous implementation shelled out to
# gstack-brain-enqueue once per file, but Windows Python cannot exec a
# bash-shebang script (the spawn fails with a fork error), so discovery
# enqueued nothing on Windows even after the path-match fix above.
# Writing the record here is platform-agnostic; the drain step
# (compute_paths_to_stage) still re-applies the skip-list + privacy filters.
if to_enqueue:
    ts = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
    written = []
    try:
        # One spool FILE per record (tmp write + atomic os.replace), matching
        # gstack-brain-enqueue's maildir contract: writers and the drain never
        # share an inode, so a parallel writer or drain can't race this.
        # Compact separators match the shim's JSON shape.
        os.makedirs(spool_dir, exist_ok=True)
        for i, (rel, key) in enumerate(to_enqueue):
            rec = json.dumps({"file": rel, "ts": ts}, separators=(",", ":"))
            tmp = os.path.join(spool_dir, f".tmp-{os.getpid()}-d{i}")
            with open(tmp, "w") as f:
                f.write(rec + "\n")
            os.replace(tmp, os.path.join(spool_dir, f"{int(time.time())}-{os.getpid()}-d{i}.json"))
            written.append((rel, key))
    except OSError:
        # Spool write failed (disk full, AV file lock). Leave the cursor
        # unadvanced for unwritten records so they are retried on the next
        # discover instead of being silently recorded as synced (which loses
        # the change until the file next changes).
        pass
    # Advance the cursor only for records actually written.
    for rel, key in written:
        new_cursor[rel] = key

save_cursor(cursor_path, new_cursor)
PYEOF
}

# -------- dispatch --------
case "${1:-}" in
  --once|"") subcmd_once ;;
  --status) subcmd_status ;;
  --skip-file) shift; subcmd_skip_file "${1:-}" ;;
  --drop-queue) shift; subcmd_drop_queue "${1:-}" ;;
  --discover-new) subcmd_discover_new ;;
  --help|-h)
    sed -n '2,18p' "$0" | sed 's/^# \{0,1\}//'
    ;;
  *)
    echo "Unknown subcommand: $1" >&2
    echo "Run: gstack-brain-sync --help" >&2
    exit 1
    ;;
esac