fix: correct all three adapters to real tool interfaces (found by live testing)

Parallel real-environment tests (graphify 0.9.23, Sourcebot v5 in Docker,
gbrain 0.42.56) surfaced real mismatches:

- graphify: build via `graphify update <dir>` (the local, no-LLM path) instead
  of `graphify <dir>` (which runs an LLM backend needing a key + network, so the
  old path wasn't actually local); query via `graphify query --graph <graph.json>`
  so it reads the indexed graph regardless of cwd; parse the real NODE/EDGE output
  (file lives at src=/at=) instead of an invented format.
- sourcebot: Sourcebot v5 gates /api/search behind auth — send
  `Authorization: Bearer <SOURCEBOT_API_KEY>`; treat 401/403 as PROVIDER_UNAVAILABLE;
  make status probe /api/search without following the login redirect.
- gbrain: degrade engine/DB init failures (e.g. pglite WASM, garrytan/gbrain#223)
  to PROVIDER_UNAVAILABLE with a one-line message instead of PROVIDER_ERROR + a raw
  stack dump; drop flags the real CLI doesn't define (`sync --strategy`,
  `search --source`); align put/delete to stdin, export to brain-wide.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

lib/code-intelligence/gbrain-adapter.ts

@@ -9,7 +9,8 @@
  * so it advertises all seven capabilities.
  */
 
-import { spawnGbrain } from "../gbrain-exec";
+import { spawnSync } from "child_process";
+import { spawnGbrain, buildGbrainEnv, NEEDS_SHELL_ON_WINDOWS } from "../gbrain-exec";
 import { ensureSourceRegistered, probeSource, sourcePageCount } from "../gbrain-sources";
 import {
   assertCapability,
@@ -89,7 +90,8 @@ export class GbrainProvider implements CodeProvider {
 
   async refresh(source: SourceRef, opts: OpOptions = {}): Promise<SourceStatus> {
     assertEgressConsent(this, opts);
-    this.#assertOk(spawnGbrain(["sync", "--strategy", "code", "--source", source.id], {
+    // `gbrain sync` has no `--strategy` flag (verified against gbrain 0.42.x --help).
+    this.#assertOk(spawnGbrain(["sync", "--source", source.id], {
       baseEnv: opts.env,
       timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS,
     }));
@@ -98,8 +100,10 @@ export class GbrainProvider implements CodeProvider {
 
   async search(query: string, opts: SearchOptions = {}): Promise<CodeSearchHit[]> {
     if (!query.trim()) return [];
+    // `gbrain search` is global and has no `--source` flag; `--limit` is real
+    // (verified against gbrain 0.42.x --help).
     const args = ["search", query];
-    if (opts.source) args.push("--source", opts.source);
+    if (opts.limit) args.push("--limit", String(opts.limit));
     const r = spawnGbrain(args, { baseEnv: opts.env, timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS });
     this.#assertOk(r);
     return parseGbrainSearch(r.stdout || "", opts.minScore ?? 0.1, opts.limit ?? 10);
@@ -129,29 +133,29 @@ export class GbrainProvider implements CodeProvider {
     }
   }
 
+  // Document ops (add/delete/export) are GBrain-only and secondary; they match
+  // gbrain's documented CLI surface (`put <slug>` reads stdin; `delete <slug>`;
+  // `export`) but could not be exercised against a live engine on the test host
+  // (pglite WASM broken, garrytan/gbrain#223), so treat them as best-effort.
   async add(doc: { slug: string; body: string }, opts: OpOptions = {}): Promise<SourceStatus> {
     assertCapability(this, "add");
     assertEgressConsent(this, opts);
-    const r = spawnGbrain(["put", doc.slug, "--body", doc.body], {
+    // `gbrain put <slug>` reads the document body from stdin.
-      baseEnv: opts.env,
+    this.#assertOk(this.#runInput(["put", doc.slug], doc.body, opts));
-      timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS,
-    });
-    this.#assertOk(r);
     return { id: doc.slug, state: "ready" };
   }
 
   async delete(slug: string, opts: OpOptions = {}): Promise<SourceStatus> {
     assertCapability(this, "delete");
-    this.#assertOk(spawnGbrain(["delete", slug, "--yes"], {
+    // stdin closed ("") so any confirmation prompt gets EOF rather than hanging.
-      baseEnv: opts.env,
+    this.#assertOk(this.#runInput(["delete", slug], "", opts));
-      timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS,
-    }));
     return { id: slug, state: "absent" };
   }
 
-  async export(source: SourceRef, opts: OpOptions = {}): Promise<string> {
+  async export(_source: SourceRef, opts: OpOptions = {}): Promise<string> {
     assertCapability(this, "export");
-    const r = spawnGbrain(["export", "--source", source.id], {
+    // `gbrain export` is brain-wide (no per-source flag); returns whatever it prints.
+    const r = spawnGbrain(["export"], {
       baseEnv: opts.env,
       timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS,
     });
@@ -159,6 +163,17 @@ export class GbrainProvider implements CodeProvider {
     return r.stdout || "";
   }
 
+  /** spawn gbrain with `input` on stdin, seeded env, Windows-shim aware. */
+  #runInput(args: string[], input: string, opts: OpOptions) {
+    return spawnSync("gbrain", args, {
+      input,
+      encoding: "utf-8",
+      timeout: opts.timeout ?? DEFAULT_TIMEOUT_MS,
+      env: buildGbrainEnv({ baseEnv: opts.env }),
+      shell: NEEDS_SHELL_ON_WINDOWS,
+    });
+  }
+
   /**
    * Throw a typed failure unless the spawn succeeded. Distinguishes a missing
    * CLI (ENOENT → PROVIDER_UNAVAILABLE, the degrade signal) from a timeout
@@ -178,21 +193,29 @@ export class GbrainProvider implements CodeProvider {
     if (r.error?.code === "ETIMEDOUT" || r.signal === "SIGTERM") {
       throw new CodeProviderError("PROVIDER_TIMEOUT", "gbrain timed out", this.id);
     }
-    if (/not configured|Cannot connect to database|config\.json/.test(stderr)) {
+    // Engine / DB / config problems are ENVIRONMENTAL — degrade to UNAVAILABLE
-      throw new CodeProviderError("PROVIDER_UNAVAILABLE", stderr || "gbrain not configured", this.id);
+    // (caller falls back to file-only), not a hard PROVIDER_ERROR with a raw dump.
+    // Covers the real case where gbrain's pglite engine fails to init its WASM
+    // runtime (garrytan/gbrain#223) as well as unreachable/unconfigured databases.
+    if (/PGLite|WASM|failed to initialize|Aborted|Cannot connect to database|not configured|config\.json|database (is )?un(reachable|available)/i.test(stderr)) {
+      throw new CodeProviderError("PROVIDER_UNAVAILABLE", firstLine(stderr) || "gbrain engine unavailable", this.id);
     }
-    throw new CodeProviderError("PROVIDER_ERROR", stderr || `gbrain exited ${r.status}`, this.id);
+    throw new CodeProviderError("PROVIDER_ERROR", firstLine(stderr) || `gbrain exited ${r.status}`, this.id);
   }
 
   #wrap(err: unknown): CodeProviderError {
     if (err instanceof CodeProviderError) return err;
     const message = err instanceof Error ? err.message : String(err);
-    if (/not on PATH|command not found/.test(message)) {
+    // Same environmental-vs-real split as #assertOk: missing CLI, or engine/DB/
-      return new CodeProviderError("PROVIDER_UNAVAILABLE", message, this.id);
+    // config problems, degrade to UNAVAILABLE so callers fall back to file-only.
+    if (/not on PATH|command not found|PGLite|WASM|failed to initialize|Aborted|Cannot connect to database|not configured|config\.json/i.test(message)) {
+      return new CodeProviderError("PROVIDER_UNAVAILABLE", firstLine(message), this.id);
     }
-    if (/not configured/.test(message)) {
+    return new CodeProviderError("PROVIDER_ERROR", firstLine(message), this.id);
-      return new CodeProviderError("PROVIDER_UNAVAILABLE", message, this.id);
-    }
-    return new CodeProviderError("PROVIDER_ERROR", message, this.id);
   }
 }
+
+/** First non-empty line, so a multi-line WASM/stack dump never reaches the user. */
+function firstLine(text: string): string {
+  return (text || "").split("\n").map((l) => l.trim()).find(Boolean) ?? "";
+}

lib/code-intelligence/graphify-adapter.ts

@@ -1,15 +1,19 @@
 /**
  * Graphify adapter — real CLI integration (github.com/Graphify-Labs/graphify).
  *
- * Graphify is a LOCAL tree-sitter knowledge graph: `graphify <dir>` builds a
+ * Graphify is a LOCAL tree-sitter knowledge graph. The genuinely local, no-LLM
- * `graphify-out/` (graph.json + report) in that dir, `graphify query "<q>"`
+ * build is `graphify update <dir>` — it writes `<dir>/graphify-out/graph.json`
- * queries it, and nothing leaves the machine (no embeddings, no network). So it
+ * with NO embeddings and NO network (verified against graphify 0.9.23). NOTE:
- * needs no egress consent and no MCP — the runtime just shells out to the CLI,
+ * the bare `graphify <dir>` build instead runs LLM semantic extraction (a gemini
- * the same shape as the GBrain adapter.
+ * backend needing an API key + network), so this adapter deliberately uses
+ * `graphify update`, which keeps the "local, no egress consent" invariant true.
  *
- * Never auto-installed: install is `pip install graphifyy && graphify install`,
+ * Query is `graphify query "<q>" --graph <dir>/graphify-out/graph.json`; the
- * a user action the picker gates on. When the CLI is absent every op throws
+ * `--graph` flag points at the built graph so search never depends on cwd.
- * PROVIDER_UNAVAILABLE and callers degrade to file-only.
+ *
+ * Never auto-installed: install is `pip install graphifyy && graphify install`
+ * (needs Python >= 3.10), a user action the picker surfaces. When the CLI is
+ * absent every op throws PROVIDER_UNAVAILABLE and callers degrade to file-only.
  *
  * Path-based, not id-based: for Graphify a source "id" IS the absolute repo path
  * (that is where `graphify-out/` lives), unlike GBrain's short source ids.
@@ -86,29 +90,28 @@ export class GraphifyProvider implements CodeProvider {
     throw new CodeProviderError("PROVIDER_ERROR", stderr || `graphify exited ${r.status}`, this.id);
   }
 
-  /** Build the graph over repo.path (local; no egress consent needed). */
+  /** Build the graph over repo.path locally (no LLM, no egress consent needed). */
   async registerSource(repo: RepoRef, opts: OpOptions = {}): Promise<SourceStatus> {
-    this.#assertOk(this.#run(["."], repo.path, opts.timeout ?? DEFAULT_TIMEOUT_MS));
+    this.#assertOk(this.#run(["update", repo.path], repo.path, opts.timeout ?? DEFAULT_TIMEOUT_MS));
     return this.status({ id: repo.path }, opts);
   }
 
-  /** Re-parse and merge changes into the existing graph. */
+  /** Re-parse and rebuild the graph (same local `graphify update` path). */
   async refresh(source: SourceRef, opts: OpOptions = {}): Promise<SourceStatus> {
-    this.#assertOk(this.#run([".", "--update"], source.id, opts.timeout ?? DEFAULT_TIMEOUT_MS));
+    this.#assertOk(this.#run(["update", source.id], source.id, opts.timeout ?? DEFAULT_TIMEOUT_MS));
     return this.status(source, opts);
   }
 
   /**
-   * `graphify query "<q>"` returns a traced answer over the graph. Its stdout is
+   * `graphify query "<q>" --graph <graph.json>` traces the graph and prints
-   * an answer, not a fixed hit schema (the CLI reference documents the command
+   * `NODE ...` / `EDGE ...` lines (plus a `Traversal:` header). We pass `--graph`
-   * but not a machine format), so we map non-empty output lines to hits
+   * explicitly so the query reads the indexed repo's graph regardless of cwd.
-   * tolerantly: a leading path-like token becomes the ref, the line the snippet.
-   * Reconcile against a live graphify if a stricter schema is needed.
    */
   async search(query: string, opts: SearchOptions = {}): Promise<CodeSearchHit[]> {
     if (!query.trim()) return [];
-    const cwd = opts.source ?? this.#root;
+    const root = opts.source ?? this.#root;
-    const r = this.#run(["query", query], cwd, opts.timeout ?? DEFAULT_TIMEOUT_MS);
+    const graphPath = join(root, OUT_DIR, GRAPH_JSON);
+    const r = this.#run(["query", query, "--graph", graphPath], root, opts.timeout ?? DEFAULT_TIMEOUT_MS);
     this.#assertOk(r);
     return parseGraphifyQuery(r.stdout || "", opts.limit ?? 10);
   }
@@ -138,19 +141,38 @@ export class GraphifyProvider implements CodeProvider {
 }
 
 /**
- * Map `graphify query` stdout to hits. Tolerant: each non-empty line becomes a
+ * Parse real `graphify query` output into hits. The format (graphify 0.9.23):
- * hit; a leading `path` or `path:line` token becomes the ref, else the whole
+ *   Traversal: BFS depth=2 | Start: ['query()'] | ... | 4 nodes found
- * line is the snippet. Exported for deterministic unit testing.
+ *   NODE query() [src=db.py loc=L4 community=login]
+ *   EDGE query() --calls [EXTRACTED context=call]--> login() at=auth.py:L8
+ * The file lives mid-line (`src=<file> loc=L<n>` on NODE, `at=<file>:L<n>` on
+ * EDGE), so the ref is `<file>:L<n>`. The `Traversal:` header and any other line
+ * are skipped. Exported for deterministic unit testing against the real format.
  */
 export function parseGraphifyQuery(stdout: string, limit: number): CodeSearchHit[] {
   const hits: CodeSearchHit[] = [];
   for (const raw of stdout.split("\n")) {
     const line = raw.trim();
-    if (!line) continue;
+    let ref: string | undefined;
-    const token = line.split(/\s+/)[0];
+    const node = line.match(/^NODE\b.*?\[src=(\S+)\s+loc=(L\d+)/);
-    const looksPath = /[/\\.]/.test(token) && !token.includes(" ");
+    const edge = line.match(/^EDGE\b.*?\bat=(\S+?):(L\d+)\b/);
-    hits.push({ ref: looksPath ? token : "graphify", snippet: line, kind: "graph-node" });
+    if (node) ref = `${node[1]}:${node[2]}`;
+    else if (edge) ref = `${edge[1]}:${edge[2]}`;
+    else continue; // skip the Traversal header and anything non-NODE/EDGE
+    hits.push({ ref, snippet: line, kind: "graph-node" });
     if (hits.length >= limit) break;
   }
   return hits;
 }
+
+/** Whether the `graphify` CLI is installed (for the picker's availability probe). */
+export function graphifyInstalled(env?: NodeJS.ProcessEnv): boolean {
+  const r = spawnSync("graphify", ["--version"], {
+    encoding: "utf-8",
+    timeout: 5_000,
+    stdio: ["ignore", "ignore", "ignore"],
+    env,
+    shell: NEEDS_SHELL_ON_WINDOWS,
+  });
+  return r.status === 0;
+}

lib/code-intelligence/sourcebot-adapter.ts

@@ -45,6 +45,12 @@ export interface SourcebotOptions {
   baseUrl?: string;
   /** Path to the server's config.json (for register_source). Defaults to SOURCEBOT_CONFIG. */
   configPath?: string;
+  /**
+   * API key for the Sourcebot REST API. Defaults to SOURCEBOT_API_KEY. Sourcebot
+   * v5 gates `/api/search` behind auth (`Authorization: Bearer <key>`); without
+   * it, `search` gets HTTP 401. Generate one in Settings -> API Keys.
+   */
+  apiKey?: string;
   /** Injectable fetch for tests. */
   fetch?: FetchLike;
   env?: NodeJS.ProcessEnv;
@@ -66,17 +72,23 @@ export class SourcebotProvider implements CodeProvider {
   readonly local: boolean;
   readonly #baseUrl: string;
   readonly #configPath?: string;
+  readonly #apiKey?: string;
   readonly #fetch: FetchLike;
 
   constructor(opts: SourcebotOptions = {}) {
     const env = opts.env ?? process.env;
     this.#baseUrl = (opts.baseUrl ?? env.SOURCEBOT_URL ?? DEFAULT_URL).replace(/\/$/, "");
     this.#configPath = opts.configPath ?? env.SOURCEBOT_CONFIG;
+    this.#apiKey = opts.apiKey ?? env.SOURCEBOT_API_KEY;
     this.#fetch = opts.fetch ?? globalThis.fetch;
     this.local = isLoopback(this.#baseUrl);
     assertRequiredCapabilities(this.id, this.capabilities);
   }
 
+  #authHeaders(): Record<string, string> {
+    return this.#apiKey ? { Authorization: `Bearer ${this.#apiKey}` } : {};
+  }
+
   has(capability: CodeProviderCapability): boolean {
     return this.capabilities.has(capability);
   }
@@ -127,8 +139,17 @@ export class SourcebotProvider implements CodeProvider {
   }
 
   async status(_source?: SourceRef, opts: OpOptions = {}): Promise<SourceStatus> {
+    // `redirect: manual` so an auth-gated server (307 -> /login) reads as
+    // not-usable instead of following to a 200 and falsely reporting "ready".
     try {
-      const res = await this.#fetchWithTimeout(this.#baseUrl, { method: "GET" }, opts.timeout ?? DEFAULT_TIMEOUT_MS);
+      const res = await this.#fetchWithTimeout(
+        `${this.#baseUrl}/api/search`,
+        { method: "POST", headers: { "Content-Type": "application/json", ...this.#authHeaders() }, body: JSON.stringify({ query: "sourcebot", matches: 1, isRegexEnabled: false }), redirect: "manual" },
+        opts.timeout ?? DEFAULT_TIMEOUT_MS,
+      );
+      if (res.status === 401 || res.status === 403) {
+        return { id: "*", state: "unknown", partial: true, detail: "reachable but not authenticated (set SOURCEBOT_API_KEY)" };
+      }
       return { id: "*", state: res.ok ? "ready" : "unknown", partial: true, detail: `HTTP ${res.status}` };
     } catch {
       return { id: "*", state: "unknown", partial: true, detail: `unreachable at ${this.#baseUrl}` };
@@ -140,12 +161,15 @@ export class SourcebotProvider implements CodeProvider {
     try {
       res = await this.#fetchWithTimeout(`${this.#baseUrl}${path}`, {
         method: "POST",
-        headers: { "Content-Type": "application/json", Accept: "application/json" },
+        headers: { "Content-Type": "application/json", Accept: "application/json", ...this.#authHeaders() },
         body: JSON.stringify(body),
       }, timeout);
     } catch (err) {
       throw new CodeProviderError("PROVIDER_UNAVAILABLE", `Sourcebot unreachable at ${this.#baseUrl}: ${(err as Error).message}`, this.id);
     }
+    if (res.status === 401 || res.status === 403) {
+      throw new CodeProviderError("PROVIDER_UNAVAILABLE", `Sourcebot requires authentication; set SOURCEBOT_API_KEY (HTTP ${res.status})`, this.id);
+    }
     if (!res.ok) throw new CodeProviderError("PROVIDER_ERROR", `Sourcebot ${path} returned HTTP ${res.status}`, this.id);
     try {
       return await res.json();