test: reset gateway in afterAll in cli-multimodal-integration — cross-file config leak

Review hardening on #3116: the file resets the gateway in beforeEach but left it configured after the last test, leaking a multimodal gateway config into whatever file runs next in its shard (same class as the adaptive-embed-batch leak fixed on master). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
fix(config): route remaining hand-rolled configureGateway callsites through buildGatewayConfig
2026-08-17 10:22:34 +00:00 · 2026-07-22 12:12:41 -07:00 · 2026-07-21 14:34:19 -07:00
15 changed files with 113 additions and 191 deletions
@@ -79,7 +79,7 @@ per-release `**vX.Y.Z:**` narration — CI enforces this
 - `docs/architecture/RETRIEVAL.md` + `docs/architecture/RETRIEVAL_MAXPOOL_INCIDENT.md` — retrieval-pipeline architecture reference + the named-thing-miss incident write-up (root cause, the five-layer fix, the eval that pins it).
 - `src/core/types.ts` extension + `src/core/operations.ts:search` + `src/core/import-file.ts` + `src/cli.ts` + `src/core/search/telemetry.ts` — the wiring layer for the retrieval cathedral. `SearchResult` gains `evidence`, `create_safety`, `title_match_boost`, `alias_hit` (all optional; evidence/create_safety reference the union types in `evidence.ts`). The `search` MCP op uses a cheap-hybrid path by default and accepts a per-call `mode` (conservative|balanced|tokenmax) honored ONLY for trusted/local callers (`resolvePerCallMode(ctx, ...)` — remote callers use the configured mode so a remote provider can't force tokenmax spend); every search path stamps evidence fail-soft. `importFromContent` projects frontmatter `aliases:` into `page_aliases` via `normalizeAliasList` + `engine.setPageAliases` so new + changed pages register aliases at ingest. `src/cli.ts` adds the `gbrain search diagnose` dispatch (lazy import) and reconciles the `search` CLI path with the cheap-hybrid op. `src/core/search/telemetry.ts` extends the rollup with the rank-1 base_score drift signal (sum/count + 3 coarse buckets, aggregate not per-query), surfaced via `gbrain search stats`, backed by migration v111's `search_telemetry` columns. Tests: `test/cli-search-dispatch.test.ts`, `test/search/per-call-mode.test.ts`, `test/search/telemetry-rank1.test.ts`, `test/search/title-boost-stage.test.ts`, `test/search/alias-hop.test.ts`, `test/search/evidence.test.ts`, `test/search/searchvector-maxpool.test.ts`, `test/search/pre-migration-failopen.test.ts`.
 - `src/commands/eval.ts` — `gbrain eval` command: single-run table + A/B config comparison. Sub-subcommand dispatch on `args[0]` routes `gbrain eval export` + `gbrain eval prune` + `gbrain eval replay` into session-capture handlers; bare `gbrain eval --qrels …` fall-through preserves the legacy IR-metrics flow. `gbrain eval cross-modal` is in the dispatch (the user-facing path is the cli.ts no-DB branch — `src/commands/eval.ts:cross-modal` only fires when callers re-enter with an existing engine).
- `src/commands/eval-cross-modal.ts` — multi-model quality gate. Three different-provider frontier models score the OUTPUT against the TASK on a 5-dim list. Verdict `pass` (exit 0) / `fail` (exit 1) / `inconclusive` (exit 2; <2/3 model successes). Reuses `src/core/ai/gateway.ts:chat()` so config/auth/aliasing comes from the gateway recipe registry — no parallel provider stack. Self-configures the gateway (`configureGateway(loadConfig() + process.env)`) since the cli.ts dispatch bypasses `connectEngine()`. Default cycles 3 in TTY, 1 in non-TTY (partial cost guardrail) via the shared `resolveCycleDefault(explicit, isTty)` in `src/core/eval/cycle-default.ts`; the cost-estimate banner appends `cycleDefaultSuffix(...)` (`for 1 cycle(s) (non-interactive default; --cycles N for more)`) when the value is the silent non-TTY fallback, so the 1-vs-3 difference isn't hidden. Receipts land at `gbrainPath('eval-receipts')/<slug>-<sha8-of-output>.json`. `--batch <jsonl> [--limit N] [--concurrent N] [--max-usd FLOAT] [--yes]` fans out cross-modal scoring across a LongMemEval-shape JSONL; mutually exclusive with `--task` (fail-fast usage error if both set); filters `kind: "by_type_summary"` rows; pre-flight cost estimate refuses if `> --max-usd` without `--yes` (default cap 5.00 USD). Semaphore-bounded fan-out via inline `runWithLimit<T>(items, limit, fn)` (exported for unit tests): max N questions in-flight × 3 model slots = ceiling of 3N parallel API calls (default `--concurrent 3` → 9). Per-question receipts land in a per-batch tempdir and are deleted at end of run; the summary receipt inlines per-question verdicts as JSON, not file paths. Exit precedence (batch-level policy, NOT inherited from aggregate.ts): ERROR > FAIL > INCONCLUSIVE > PASS. DI seam: `runEvalCrossModal(args, opts?: {runEval?: typeof runEval})` mirrors `runEvalLongMemEval(args, {client?})`; tests pass `opts.runEval` to bypass real LLM calls AND the gateway availability check. Pinned by `test/eval-cross-modal-batch.test.ts`.
+- `src/commands/eval-cross-modal.ts` — multi-model quality gate. Three different-provider frontier models score the OUTPUT against the TASK on a 5-dim list. Verdict `pass` (exit 0) / `fail` (exit 1) / `inconclusive` (exit 2; <2/3 model successes). Reuses `src/core/ai/gateway.ts:chat()` so config/auth/aliasing comes from the gateway recipe registry — no parallel provider stack. Self-configures the gateway via `buildGatewayConfig(loadConfig() ?? {})` since the cli.ts dispatch bypasses `connectEngine()`, so file-plane keys, env base URLs, and provider_chat_options follow the same adapter path as runtime. Default cycles 3 in TTY, 1 in non-TTY (partial cost guardrail) via the shared `resolveCycleDefault(explicit, isTty)` in `src/core/eval/cycle-default.ts`; the cost-estimate banner appends `cycleDefaultSuffix(...)` (`for 1 cycle(s) (non-interactive default; --cycles N for more)`) when the value is the silent non-TTY fallback, so the 1-vs-3 difference isn't hidden. Receipts land at `gbrainPath('eval-receipts')/<slug>-<sha8-of-output>.json`. `--batch <jsonl> [--limit N] [--concurrent N] [--max-usd FLOAT] [--yes]` fans out cross-modal scoring across a LongMemEval-shape JSONL; mutually exclusive with `--task` (fail-fast usage error if both set); filters `kind: "by_type_summary"` rows; pre-flight cost estimate refuses if `> --max-usd` without `--yes` (default cap 5.00 USD). Semaphore-bounded fan-out via inline `runWithLimit<T>(items, limit, fn)` (exported for unit tests): max N questions in-flight × 3 model slots = ceiling of 3N parallel API calls (default `--concurrent 3` → 9). Per-question receipts land in a per-batch tempdir and are deleted at end of run; the summary receipt inlines per-question verdicts as JSON, not file paths. Exit precedence (batch-level policy, NOT inherited from aggregate.ts): ERROR > FAIL > INCONCLUSIVE > PASS. DI seam: `runEvalCrossModal(args, opts?: {runEval?: typeof runEval})` mirrors `runEvalLongMemEval(args, {client?})`; tests pass `opts.runEval` to bypass real LLM calls AND the gateway availability check. Pinned by `test/eval-cross-modal-batch.test.ts`.
 - `src/core/eval/cycle-default.ts` — single source of truth for the eval cycle-count default. Exports `DEFAULT_CYCLES_TTY = 3`, `DEFAULT_CYCLES_NONTTY = 1`, `resolveCycleDefault(explicit, isTty): {cycles, usedNonTtyDefault}`, and `cycleDefaultSuffix(r)` (returns ` (non-interactive default; --cycles N for more)` only when the non-TTY default was applied, else `''`). Consumed by `eval-cross-modal.ts`, `eval-takes-quality.ts` (run + regress), and `takes-quality-eval/runner.ts` (core uses only the constant — library stays TTY-agnostic; the CLI owns the TTY=3 upgrade + banner annotation). `eval-suspected-contradictions.ts` applies the same transparency to its `$5`/`$1` budget default via a `budgetUsdExplicit` flag (the budget is overwritten in-place so explicitness can't be inferred post-hoc). Not shared with `resolveWorkersWithClamp` (different domain, no engine, no dedup). Pinned by `test/eval/cycle-default.test.ts`, `test/eval-suspected-contradictions-budget-default.test.ts`.
 - `src/core/cross-modal-eval/json-repair.ts` — `parseModelJSON(raw)` named export with a 4-strategy fallback chain (direct parse → fence-strip → trailing-comma + single-quote + embedded-newline repair → regex nuclear option). Adversarial input throws rather than fabricating scores — the aggregator treats a throw as "this model contributed nothing this cycle" so the gate stays correct at >=2/3 successes.
 - `src/core/cross-modal-eval/aggregate.ts` — pure verdict logic. Pass criterion: `(successes >= 2) AND (every dim mean >= 7) AND (every dim min across models >= 5)`. Inconclusive when <2/3 models returned parseable scores (regression guard for the v1 `Object.values({}).every(...) === true` empty-array PASS bug).
@@ -138,7 +138,7 @@ per-release `**vX.Y.Z:**` narration — CI enforces this
 - `src/core/minions/queue.ts` extension — `MinionQueue.add()` rejects `subagent` jobs whose `data.model` resolves via `isAnthropicProvider()` to a non-Anthropic provider. Lazy-imports `model-config.ts` to avoid pulling engine types into queue's eager-load surface. Layer 1 of the three-layer subagent provider enforcement (layers 2+3: `model-config.ts:enforceSubagentAnthropic` runtime fallback + `src/commands/doctor.ts` `subagent_provider` check). Pinned by `test/agent-cli.test.ts`.
 - `src/commands/models.ts` — `gbrain models [--json]` read-only routing dashboard: prints tier defaults (`utility`/`reasoning`/`deep`/`subagent`), the resolved value for each (re-walking the resolution chain), every per-task override (11 `PER_TASK_KEYS`: `models.dream.synthesize`, `models.dream.patterns`, `models.drift`, `models.auto_think`, `models.think`, `models.subagent`, `facts.extraction_model`, `models.eval.longmemeval`, `models.expansion`, `models.chat`, `models.dream.synthesize_verdict`), the alias map, and a source-of-truth column (`default` / `config: <key>` / `env: <VAR>`). `gbrain models doctor [--skip=<provider>] [--json]` fires a 1-token `gateway.chat()` probe against each configured chat + expansion model and classifies failures into `{model_not_found, auth, rate_limit, network, unknown}`. Wired into `cli.ts` dispatch + `CLI_ONLY` set. A zero-token `embedding_config` probe runs FIRST, before any chat/expansion probes spend money: `probeEmbeddingConfig()` reads `getEmbeddingModel()` + `getEmbeddingDimensions()` and (for Voyage flexible-dim models) checks `isValidVoyageOutputDim(dims)` against `VOYAGE_VALID_OUTPUT_DIMS`. `ProbeStatus` variant `'config'` + optional `fix?: string` on `ProbeResult` surface a paste-ready `gbrain config set ...` line in human + JSON output; touchpoint label `'embedding_config'` joins `'chat'` and `'expansion'`.
 - `src/core/init-embed-check.ts` — embedding-key validation at `gbrain init`. `runInitEmbedCheck(opts)` runs a config-only `diagnoseEmbedding` (catches a missing key for ANY provider) plus a best-effort `liveTestEmbed` (1-token `gateway.embed(['probe'], {inputType:'query', abortSignal})`, 5s `AbortController` timeout, never throws — catches an invalid/expired key). Loud warning to stderr; init still exits 0 (`--no-embedding` is the deferred-setup escape; `--skip-embed-check` / `GBRAIN_INIT_SKIP_EMBED_CHECK=1` skip the check). Builds the effective env (`process.env` + file-plane `openai/anthropic/zeroentropy_api_key` from `loadConfigFileOnly()` + `opts.apiKey`) and configures the gateway via `buildGatewayConfig` before diagnose/probe, so the check sees the same keys AND provider base URLs runtime will (no false "missing key" for config.json-keyed users; the probe hits the right endpoint). Init-specific warning text names `--no-embedding` / `--skip-embed-check`, not the sync-flavored `--no-embed`. Wired into `initPGLite` + `initPostgres` in `src/commands/init.ts`, with the result added to the `--json` envelope as `embedding_check {ok, reason?, live_ok?}`. Pinned by `test/init-embed-check.test.ts` (hermetic via the gateway embed-transport seam + `withEnv`).
- `src/core/ai/build-gateway-config.ts` — `buildGatewayConfig(c: GBrainConfig): AIGatewayConfig`, extracted from `src/cli.ts` (which re-exports it for back-compat). Lets core modules (`init-embed-check.ts`) reuse it without importing the CLI entrypoint. Single owner of folding file-plane API keys (openai/anthropic/zeroentropy) into the gateway env and threading local-server `*_BASE_URL` env vars into base_urls. `process.env` wins EXCEPT empty-string / undefined values are dropped before the merge, so an injected empty `ANTHROPIC_API_KEY=''` (Claude Code neuters subprocess LLM calls this way) can't clobber a valid config-plane key; `'0'` / `'false'` are preserved. Pinned by `test/ai/build-gateway-config.test.ts`.
+- `src/core/ai/build-gateway-config.ts` — `buildGatewayConfig(c: GBrainConfig): AIGatewayConfig`, extracted from `src/cli.ts` (which re-exports it for back-compat). Single owner of translating stored config into gateway config — consumed by CLI runtime, init (`gbrain init`'s three configureGateway sites), `init-embed-check.ts`, the eval commands (cross-modal, takes-quality), provider diagnostics, and the in-process migration path. Folds file-plane API keys (openai/anthropic/zeroentropy/openrouter) into the gateway env and threads local-server `*_BASE_URL` env vars into base_urls; caller-provided `provider_base_urls` config wins over env base URLs. `process.env` wins EXCEPT empty-string / undefined values are dropped before the merge, so an injected empty `ANTHROPIC_API_KEY=''` (Claude Code neuters subprocess LLM calls this way) can't clobber a valid config-plane key; `'0'` / `'false'` are preserved. Pinned by `test/ai/build-gateway-config.test.ts`.
 - `src/commands/doctor.ts` extension — `subagent_provider` check (layer 3 of 3). Warns when `models.tier.subagent` is explicitly set non-Anthropic (message names the bad value + paste-ready fix `gbrain config set models.tier.subagent anthropic:claude-sonnet-4-6`); also warns when `models.default` would sneak `subagent` into a non-Anthropic provider via tier inheritance. OK when subagent tier resolves to Anthropic. Tests in `test/doctor.test.ts`.
 - `src/core/skill-trigger-index.ts` — Shared loader that unions per-skill SKILL.md frontmatter `triggers:` with curated RESOLVER.md / AGENTS.md rows from `skillsDir` AND the parent dir (preserves the OpenClaw workspace-root layout). UNION semantics: explicit RESOLVER.md rows ADD to frontmatter triggers (don't replace). Dedup keyed on `(skillPath, trigger.trim().toLowerCase())`. Three consumers fold through this primitive — `checkResolvable`, `runRoutingEvalCli`, `mounts-cache.composeResolvers` — so fixing frontmatter reaches all of them. Exports `loadSkillTriggerIndex(skillsDir): SkillTriggerEntry[]`, `entriesToResolverContent(entries): string` (synthesizes a markdown-table resolver string for `runRoutingEval`'s string-content API), `findPrimaryResolverPath(skillsDir): string | null`, the `FRONTMATTER_SECTION` constant, and `_resetWarnedSkillsForTests`. Skip rules: non-directory entries, `_*`/`.*` prefixes, `conventions/`+`migrations/` subdirs, skills with no `SKILL.md` (deprecated `install/` graceful-skipped), no `triggers:` array, or malformed YAML (warn-once + skip). Reuses `parseSkillFrontmatter` from `src/core/skill-frontmatter.ts` (regex-based, not full YAML). Pinned by `test/skill-trigger-index.test.ts` (18 hermetic cases). CI gate `bun run check:resolver` (= `bun src/cli.ts check-resolvable --strict --skills-dir skills/`) wired into `bun run verify`.
 - `src/core/skill-catalog.ts` — host-repo skill catalog backing the MCP `list_skills` / `get_skill` ops. Lets a thin MCP client (Codex desktop, Claude Code, Claude Cowork, Perplexity) DISCOVER + FOLLOW the agent repo's fat-markdown skills over `gbrain serve` — a skill is prose, so "using" one = fetching its body then calling the gbrain MCP tools the server already exposes. Read-scope, NOT localOnly (defensible only via the full mitigation stack): (1) **publish gate** — `assertPublishEnabled(ctx, publishSkills)`; remote callers require `mcp.publish_skills === true`, default-OFF so an upgrade never silently grants existing read tokens host-skill read; local callers (`ctx.remote === false`) always pass. (2) **path confinement** — `assertSkillNameShape` rejects separators/`..`/null/space before any FS access; the client `name` is a manifest LOOKUP KEY (via `loadOrDeriveManifest`), never a raw path segment; `confineManifestPath` does realpath + relative-containment + `SKILL.md`-regular-file check on EVERY entry (defeats poisoned manifest.json `path`, symlink/`..` escape). (3) **frontmatter allowlist** — `GetSkillResult.frontmatter` projects a safe subset; private `writes_to` + `sources` dropped. (4) **prose-only + 256KB cap** (`MAX_SKILL_MD_BYTES`, env `GBRAIN_MAX_SKILL_MD_BYTES`), size-checked twice (statSync + UTF-8 byte length). (5) **no install_path serve for remote** — remote callers use `autoDetectSkillsDir` (no install-path tier) so a hosted gbrain with no agent repo returns `storage_error`; local callers use `autoDetectSkillsDirReadOnly`. (6) MCP rate-limiter caps call rate. Config reads honor BOTH planes: `readMcpPublishSkills` / `readMcpSkillsDir` prefer the DB plane (`engine.getConfig`) over the file plane (`ctx.config.mcp`). Tool-honesty: `crossReferenceTools(declared, ctx)` splits a skill's declared `tools:` into `usable_tools` vs `unavailable_tools`; `buildSkillCatalog`'s `instructions` envelope (`SKILL_CATALOG_INSTRUCTIONS`) carries the "these are prose, follow-then-call-tools" protocol. Skills are host-filesystem repo-global — `sourceScopeOpts(ctx)` / `ctx.brainId` deliberately do NOT apply. `buildSkillCatalog` is resilient (one malformed/escaping skill is skipped, never throws). Config keys in `src/core/config.ts`: `GBrainConfig.mcp?: { publish_skills?, skills_dir? }` + `KNOWN_CONFIG_KEYS` entries `mcp.publish_skills`/`mcp.publish_skills_prompted`/`mcp.skills_dir` + `mcp.` prefix in `KNOWN_CONFIG_KEY_PREFIXES`. `src/commands/init.ts` writes `config.mcp = { publish_skills: true, ... }` for new installs (existing config wins on re-init). `src/commands/upgrade.ts:runPostUpgrade` adds a one-time consent prompt (gated by `mcp.publish_skills_prompted`; existing installs stay OFF until owner opts in). Two ops register in `src/core/operations.ts` (`list_skills` with optional `section` filter + `cliHints:{name:'skills'}`; `get_skill` taking `name` + `cliHints:{name:'skill', positional:['name']}`) and dynamically import this module to avoid the import cycle (skill-catalog statically imports the `operations` array). Descriptions in `src/core/operations-descriptions.ts` (`LIST_SKILLS_DESCRIPTION`, `GET_SKILL_DESCRIPTION`, `SKILL_CATALOG_INSTRUCTIONS`, `SKILL_CLIENT_GUIDANCE`), pinned by `test/operations-descriptions.test.ts`. CLI: `gbrain skills` / `gbrain skill <name>`. Pinned by `test/skill-catalog.test.ts`, `test/skill-catalog-security.test.ts` (path-confinement / poisoned-manifest / symlink-escape), `test/skill-catalog-transports.test.ts` (publish-gate + remote-vs-local) over `test/fixtures/skill-catalog/`.
@@ -186,18 +186,6 @@ export function isSourceStale(src: SourceRow, now = Date.now(), floorMin = FULL_
  return ageMin >= floorMin;
 }

-/**
- * #2060: count sources past the per-source cycle freshness floor. Consumed
- * by autopilot's dispatch decision — a stale source forces the fanout path
- * even when the doctor plan is small (score 70–94, plan ≤ 3, est < 300s),
- * so targeted mode can't leave cycle_freshness stale indefinitely.
- * dispatchPerSource's own throttles (skipped_fresh / fanoutMax / failure
- * cooldown) bound the resulting work.
- */
-export function countStaleSources(sources: SourceRow[], now = Date.now(), floorMin = FULL_CYCLE_FLOOR_MIN): number {
-  return sources.filter((s) => isSourceStale(s, now, floorMin)).length;
-}
-
 /**
 * Most recent SUCCESSFUL cycle for a source. Prefers `last_source_cycle_at`
 * (per-source phases, written by the split cycle) and falls back to the legacy
@@ -901,27 +901,13 @@ export async function runAutopilot(engine: BrainEngine, args: string[]) {
        const FULL_CYCLE_FLOOR_MIN = 60;
        const minutesSinceLastFull = (Date.now() - lastFullCycleAt) / 60000;

-        // #2060: stale per-source cycle freshness is a dispatch input. Without
-        // it, a brain sitting at score 70–94 with a small targeted plan (≤3
-        // steps, <300s) stays in targeted mode indefinitely and no per-source
-        // cycle is ever dispatched — cycle_freshness never advances. A stale
-        // source forces the fanout path; dispatchPerSource's throttles
-        // (skipped_fresh / fanoutMax / failure cooldown) bound the work.
-        // Fail-open to 0: a read failure must not block dispatch.
-        let staleCycleSources = 0;
-        try {
-          const { countStaleSources } = await import('./autopilot-fanout.ts');
-          staleCycleSources = countStaleSources(await engine.listAllSources({ localPathOnly: true }));
-        } catch { /* fail-open: freshness is a dispatch hint, not a gate */ }
-
        const shouldFullCycle =
          (score >= 95 && plan.length === 0 && minutesSinceLastFull >= FULL_CYCLE_FLOOR_MIN) ||
          plan.length > 3 ||
          estTotal >= 300 ||
-          score < 70 ||
-          staleCycleSources > 0;
+          score < 70;

-        const shouldSleep = score >= 95 && plan.length === 0 && minutesSinceLastFull < FULL_CYCLE_FLOOR_MIN && staleCycleSources === 0;
+        const shouldSleep = score >= 95 && plan.length === 0 && minutesSinceLastFull < FULL_CYCLE_FLOOR_MIN;

        if (shouldSleep) {
          if (jsonMode) {
@@ -21,7 +21,8 @@ import { join } from 'path';
 import { tmpdir } from 'os';
 import { createHash } from 'crypto';

-import { gbrainPath, loadConfig } from '../core/config.ts';
+import { gbrainPath, loadConfig, type GBrainConfig } from '../core/config.ts';
+import { buildGatewayConfig } from '../core/ai/build-gateway-config.ts';
 import { configureGateway, isAvailable } from '../core/ai/gateway.ts';
 import { runWithLimit } from '../core/worker-pool.ts';
 import { resolveCycleDefault, cycleDefaultSuffix } from '../core/eval/cycle-default.ts';
@@ -264,32 +265,11 @@ function isTTY(): boolean {
 * Returns true on success; false (and prints a hint) when no config is found.
 */
 function configureGatewayForCli(): boolean {
+  // Route through buildGatewayConfig (the single adapter seam) so file-plane
+  // API keys, env base URLs, and provider_chat_options follow the same
+  // precedence as the runtime path. No config file is fine — env alone serves.
  const config = loadConfig();
-  if (!config) {
-    // No config file is fine for the eval command — env vars alone may serve.
-    // We still call configureGateway so gateway recipes can read the env map.
-    configureGateway({
-      embedding_model: undefined,
-      embedding_dimensions: undefined,
-      expansion_model: undefined,
-      chat_model: undefined,
-      chat_fallback_chain: undefined,
-      base_urls: undefined,
-      provider_chat_options: undefined,
-      env: { ...process.env },
-    });
-    return true;
-  }
-  configureGateway({
-    embedding_model: config.embedding_model,
-    embedding_dimensions: config.embedding_dimensions,
-    expansion_model: config.expansion_model,
-    chat_model: config.chat_model,
-    chat_fallback_chain: config.chat_fallback_chain,
-    base_urls: config.provider_base_urls,
-    provider_chat_options: config.provider_chat_options,
-    env: { ...process.env },
-  });
+  configureGateway(buildGatewayConfig(config ?? ({} as GBrainConfig)));
  return true;
 }

@@ -21,7 +21,8 @@
 */
 import type { BrainEngine } from '../core/engine.ts';
 import { configureGateway } from '../core/ai/gateway.ts';
-import { loadConfig } from '../core/config.ts';
+import { buildGatewayConfig } from '../core/ai/build-gateway-config.ts';
+import { loadConfig, type GBrainConfig } from '../core/config.ts';
 import { runEval, DEFAULT_MODEL_PANEL } from '../core/takes-quality-eval/runner.ts';
 import { resolveCycleDefault, cycleDefaultSuffix } from '../core/eval/cycle-default.ts';
 import { writeReceipt } from '../core/takes-quality-eval/receipt-write.ts';
@@ -127,8 +128,12 @@ export async function runReplayNoBrain(argv: string[]): Promise<number> {
 export async function runEvalTakesQuality(engine: BrainEngine, args: string[]): Promise<void> {
  // Self-configure the AI gateway (mirrors eval-cross-modal pattern). The
  // gateway needs config.ai_gateway + env vars; configureGateway reads both.
+  // Route through buildGatewayConfig: the old `{ ...cfg, ...process.env }`
+  // spread never populated the gateway's `env` field (the gateway NEVER reads
+  // process.env at call time), so availability checks saw no keys at all and
+  // file-plane API keys / provider base URLs were dropped.
  const cfg = loadConfig();
-  configureGateway({ ...cfg, ...(process.env as Record<string, string>) } as any);
+  configureGateway(buildGatewayConfig(cfg ?? ({} as GBrainConfig)));

  const { subcmd, argv, json } = parseSubcmd(args);

@@ -7,6 +7,7 @@ import { homedir } from 'os';
 const __filename = fileURLToPath(import.meta.url);
 const __dirname = dirname(__filename);
 import { saveConfig, loadConfig, loadConfigFileOnly, toEngineConfig, gbrainPath, configPath, isThinClient, effectiveEnvDatabaseUrl, type GBrainConfig } from '../core/config.ts';
+import { buildGatewayConfig } from '../core/ai/build-gateway-config.ts';
 import { createEngine } from '../core/engine-factory.ts';
 import { discoverOAuth, mintClientCredentialsToken, smokeTestMcp } from '../core/remote-mcp-probe.ts';
 import { runInitEmbedCheck } from '../core/init-embed-check.ts';
@@ -722,7 +723,8 @@ async function configureGatewayWithMergedPrecedence(
  // pollutes config.json.
  const envOverlay = loadConfig() ?? ({} as GBrainConfig);

-  const merged = {
+  const merged: GBrainConfig = {
+    ...envOverlay,
    embedding_model: aiOpts?.embedding_model ?? envOverlay.embedding_model ?? existingFile.embedding_model,
    embedding_dimensions: aiOpts?.embedding_dimensions ?? envOverlay.embedding_dimensions ?? existingFile.embedding_dimensions,
    expansion_model: aiOpts?.expansion_model ?? envOverlay.expansion_model ?? existingFile.expansion_model,
@@ -730,13 +732,9 @@ async function configureGatewayWithMergedPrecedence(
  };

  const { configureGateway, getEmbeddingModel, getEmbeddingDimensions, getExpansionModel, getChatModel } = await import('../core/ai/gateway.ts');
-  configureGateway({
-    embedding_model: merged.embedding_model,
-    embedding_dimensions: merged.embedding_dimensions,
-    expansion_model: merged.expansion_model,
-    chat_model: merged.chat_model,
-    env: { ...process.env },
-  });
+  // buildGatewayConfig (the single adapter seam) so file-plane API keys and
+  // env base URLs reach the gateway — a hand-rolled config here dropped them.
+  configureGateway(buildGatewayConfig(merged));

  // Read back resolved values — gateway applies internal defaults for unset
  // fields, so these are the values that actually shaped the schema.
@@ -831,13 +829,13 @@ async function initPGLite(opts: {
  // resolveAIOptions above: CLI flags > env vars > existing file > gateway
  // defaults.
  const { configureGateway } = await import('../core/ai/gateway.ts');
-  configureGateway({
+  configureGateway(buildGatewayConfig({
+    ...(loadConfig() ?? ({} as GBrainConfig)),
    embedding_model: resolvedModel ?? opts.aiOpts?.embedding_model,
    embedding_dimensions: resolvedDim ?? opts.aiOpts?.embedding_dimensions,
    expansion_model: opts.aiOpts?.expansion_model,
    chat_model: opts.aiOpts?.chat_model,
-    env: { ...process.env },
-  });
+  } as GBrainConfig));
  if (resolvedModel) console.log(`  Embedding: ${resolvedModel} (${resolvedDim}d)`);
  if (opts.aiOpts?.expansion_model) console.log(`  Expansion: ${opts.aiOpts.expansion_model}`);
  if (opts.aiOpts?.chat_model) console.log(`  Chat: ${opts.aiOpts.chat_model}`);
@@ -1046,13 +1044,13 @@ async function initPostgres(opts: {

  // T6: unconditional configureGateway BEFORE initSchema.
  const { configureGateway } = await import('../core/ai/gateway.ts');
-  configureGateway({
+  configureGateway(buildGatewayConfig({
+    ...(loadConfig() ?? ({} as GBrainConfig)),
    embedding_model: resolvedModel ?? opts.aiOpts?.embedding_model,
    embedding_dimensions: resolvedDim ?? opts.aiOpts?.embedding_dimensions,
    expansion_model: opts.aiOpts?.expansion_model,
    chat_model: opts.aiOpts?.chat_model,
-    env: { ...process.env },
-  });
+  } as GBrainConfig));
  if (resolvedModel) console.log(`  Embedding: ${resolvedModel} (${resolvedDim}d)`);
  if (opts.aiOpts?.expansion_model) console.log(`  Expansion: ${opts.aiOpts.expansion_model}`);
  if (opts.aiOpts?.chat_model) console.log(`  Chat: ${opts.aiOpts.chat_model}`);
@@ -66,15 +66,11 @@ export async function runMigrateOnlyCore(opts?: { timeoutMs?: number }): Promise

  // configureGateway BEFORE initSchema (init.ts B.3): a schema bump on a brain
  // whose file config is missing embedding fields must not fall through to
-  // stale hardcoded fallbacks. loadConfig already merged env; propagate it.
+  // stale hardcoded fallbacks. Route through buildGatewayConfig so file-plane
+  // API keys and provider base URLs follow the same precedence as runtime.
  const { configureGateway } = await import('../../core/ai/gateway.ts');
-  configureGateway({
-    embedding_model: config.embedding_model,
-    embedding_dimensions: config.embedding_dimensions,
-    expansion_model: config.expansion_model,
-    chat_model: config.chat_model,
-    env: { ...process.env },
-  });
+  const { buildGatewayConfig } = await import('../../core/ai/build-gateway-config.ts');
+  configureGateway(buildGatewayConfig(config));

  const timeoutMs = opts?.timeoutMs ?? MIGRATE_ONLY_TIMEOUT_MS;
  const engine = await createEngine(toEngineConfig(config));
@@ -12,8 +12,8 @@ import { parseModelId } from './ai/model-resolver.ts';
 * `resolveKey` closure without re-parsing recipes.
 *
 * Only OPENAI_API_KEY and ZEROENTROPY_API_KEY appear here because those are the
- * only embedding keys `buildGatewayConfig` (src/cli.ts) folds from config into
- * the gateway env. VOYAGE_API_KEY / GOOGLE_GENERATIVE_AI_API_KEY are deliberately
+ * only embedding keys `buildGatewayConfig` (src/core/ai/build-gateway-config.ts)
+ * folds from config into the gateway env. VOYAGE_API_KEY / GOOGLE_GENERATIVE_AI_API_KEY are deliberately
 * absent: their config fields are NOT threaded to the gateway today, so the
 * producer closures fall through to checking `process.env` ONLY for them. That
 * matches what the gateway can actually use (the recipes read those keys from
@@ -854,10 +854,7 @@ interface SyncPhaseResult extends PhaseResult {
 /**
 * Resolve the source id for a brain directory by looking up the sources
 * table. Returns undefined when no registered source matches (falls back
- * to pre-v0.18 global config.sync.* keys) OR when MORE than one source
- * claims the path — an ambiguous match must not scope phases or stamp
- * last_full_cycle_at for an arbitrarily-picked source (the "freshness
- * stamp that lies" this resolution exists to prevent).
+ * to pre-v0.18 global config.sync.* keys).
 */
 async function resolveSourceForDir(
  engine: BrainEngine,
@@ -868,10 +865,10 @@ async function resolveSourceForDir(
  if (brainDir === null) return undefined;
  try {
    const rows = await engine.executeRaw<{ id: string }>(
-      `SELECT id FROM sources WHERE local_path = $1 LIMIT 2`,
+      `SELECT id FROM sources WHERE local_path = $1 LIMIT 1`,
      [brainDir],
    );
-    return rows.length === 1 ? rows[0]!.id : undefined;
+    return rows[0]?.id;
  } catch {
    // sources table might not exist on very old brains — fall through.
    return undefined;
@@ -2368,23 +2365,17 @@ export async function runCycle(
  }

  // v0.38 (codex r1 P0-5): persist per-source cycle completion timestamp
-  // when the cycle ran successfully against a resolvable source. Read by
-  // autopilot's per-source freshness gate next tick.
-  //
-  // #1993: keyed off `cycleSourceId` (opts.sourceId ?? the source resolved
-  // from brainDir) — the SAME id the cycle locked + scoped its phases to —
-  // NOT raw opts.sourceId. The autopilot's inline cycle sets brainDir but
-  // passes no explicit sourceId, so keying off opts.sourceId alone never
-  // advanced last_full_cycle_at and cycle_freshness stayed stale even while
-  // the autopilot cycled every interval. Skipped when:
-  //   - no source resolves (engine null, or no checkout AND no opts.sourceId)
+  // when the cycle ran successfully against an explicit source. Read by
+  // autopilot's per-source freshness gate next tick. Skipped when:
+  //   - opts.sourceId is unset (legacy callers — autopilot still here)
+  //   - engine is null (no-DB path)
  //   - status is 'failed' or 'skipped' (don't mark a non-run as fresh)
  //   - dryRun (writes are out of scope)
  //
  // Best-effort: a write failure does NOT change the CycleReport status.
  // The cost of writing the wrong timestamp post-failure is higher than
  // the cost of missing a successful write (next cycle will redo work).
-  if (cycleSourceId && engine && !dryRun && !aborted && (status === 'ok' || status === 'clean' || status === 'partial')) {
+  if (opts.sourceId && engine && !dryRun && !aborted && (status === 'ok' || status === 'clean' || status === 'partial')) {
    try {
      const nowIso = new Date().toISOString();
      // #2194 fix #3 (the cycle split): `last_source_cycle_at` is the NEW gate
@@ -2394,13 +2385,13 @@ export async function runCycle(
      // phases (those gate on autopilot.last_global_at), so writing it on a
      // source-only cycle does not re-introduce the freshness poisoning codex
      // flagged in the rejected skip-based design.
-      await engine.updateSourceConfig(cycleSourceId, {
+      await engine.updateSourceConfig(opts.sourceId, {
        last_source_cycle_at: nowIso,
        last_full_cycle_at: nowIso,
      });
    } catch (e) {
      // Best-effort; cycle already succeeded by the time we get here.
-      console.warn(`[cycle] failed to write last_source_cycle_at for source ${cycleSourceId}: ${e instanceof Error ? e.message : String(e)}`);
+      console.warn(`[cycle] failed to write last_source_cycle_at for source ${opts.sourceId}: ${e instanceof Error ? e.message : String(e)}`);
    }
  }

@@ -25,6 +25,7 @@ import { withEnv } from '../helpers/with-env.ts';

 const PASSTHROUGHS: Array<{ envVar: string; recipeId: string }> = [
  { envVar: 'LLAMA_SERVER_BASE_URL', recipeId: 'llama-server' },
+  { envVar: 'LLAMA_SERVER_RERANKER_BASE_URL', recipeId: 'llama-server-reranker' },
  { envVar: 'OLLAMA_BASE_URL', recipeId: 'ollama' },
  { envVar: 'LMSTUDIO_BASE_URL', recipeId: 'lmstudio' },
  { envVar: 'LITELLM_BASE_URL', recipeId: 'litellm' },
@@ -54,20 +54,6 @@ describe('autopilot.ts ↔ dispatchPerSource wiring', () => {
    expect(AUTOPILOT_SRC).toMatch(/lastFullCycleAt\s*=\s*Date\.now\(\)/);
  });

-  test('stale per-source cycle freshness is a shouldFullCycle input (#2060)', () => {
-    // Targeted mode (score 70–94, plan ≤3, est <300s) must not be able to
-    // starve per-source cycle dispatch: a stale source (per countStaleSources
-    // over listAllSources) forces the fanout path, and the sleep gate must
-    // not fire while stale sources exist. Without these terms, cycle
-    // freshness never advances for a brain that always lands in targeted mode.
-    expect(AUTOPILOT_SRC).toMatch(/countStaleSources/);
-    const fullCycleDeclIdx = AUTOPILOT_SRC.indexOf('const shouldFullCycle');
-    expect(fullCycleDeclIdx).toBeGreaterThan(-1);
-    const decl = AUTOPILOT_SRC.slice(fullCycleDeclIdx, fullCycleDeclIdx + 700);
-    expect(decl).toMatch(/staleCycleSources\s*>\s*0/);
-    expect(decl).toMatch(/const shouldSleep[^;]*staleCycleSources\s*===\s*0/);
-  });
-
  test('does NOT regress to the single-job dispatch on the full-cycle path', () => {
    // Pre-PR: the shouldFullCycle branch did:
    //   const job = await queue.add('autopilot-cycle', { repoPath }, {
@@ -14,7 +14,6 @@ import { describe, test, expect } from 'bun:test';
 import {
  readLastFullCycleAt,
  isSourceStale,
-  countStaleSources,
  selectSourcesForDispatch,
  resolveFanoutMax,
  dispatchPerSource,
@@ -75,23 +74,6 @@ describe('isSourceStale', () => {
  });
 });

-describe('countStaleSources (#2060 dispatch-decision input)', () => {
-  const NOW = Date.parse('2026-05-22T12:00:00.000Z');
-  test('counts never-cycled + past-floor sources, ignores fresh', () => {
-    const sources = [
-      src('never-cycled'), // stale (null)
-      src('old', new Date(NOW - 2 * 60 * 60_000).toISOString()), // stale (2h)
-      src('fresh', new Date(NOW - 30 * 60_000).toISOString()), // fresh (30min)
-    ];
-    expect(countStaleSources(sources, NOW)).toBe(2);
-  });
-  test('returns 0 for all-fresh and for empty list', () => {
-    const fresh = src('a', new Date(NOW - 10 * 60_000).toISOString());
-    expect(countStaleSources([fresh], NOW)).toBe(0);
-    expect(countStaleSources([], NOW)).toBe(0);
-  });
-});
-
 describe('selectSourcesForDispatch', () => {
  const NOW = Date.parse('2026-05-22T12:00:00.000Z');
  const fresh = (id: string, agoMin: number) =>
@@ -11,32 +11,13 @@
 import { afterAll, beforeAll, beforeEach, describe, expect, test } from 'bun:test';
 import { PGLiteEngine } from '../src/core/pglite-engine.ts';
 import { loadConfigWithEngine, type GBrainConfig } from '../src/core/config.ts';
+import { buildGatewayConfig } from '../src/core/ai/build-gateway-config.ts';
 import {
  configureGateway,
  getEmbeddingModel,
  getMultimodalModel,
  resetGateway,
 } from '../src/core/ai/gateway.ts';
-import type { AIGatewayConfig } from '../src/core/ai/types.ts';
-
-// Mirror the cli.ts buildGatewayConfig helper exactly. Keeping a copy here
-// (instead of exporting from cli.ts) is intentional: the test asserts the
-// shape of the contract, not the helper's identity. If cli.ts drifts, the
-// e2e behavior these tests care about (DB-set value lands in gateway) still
-// holds, but a helper-shape test would also catch the drift in PR review.
-function buildGatewayConfig(c: GBrainConfig): AIGatewayConfig {
-  return {
-    embedding_model: c.embedding_model,
-    embedding_dimensions: c.embedding_dimensions,
-    embedding_multimodal_model: c.embedding_multimodal_model,
-    expansion_model: c.expansion_model,
-    chat_model: c.chat_model,
-    chat_fallback_chain: c.chat_fallback_chain,
-    base_urls: c.provider_base_urls,
-    provider_chat_options: c.provider_chat_options,
-    env: { ...process.env },
-  };
-}

 let engine: PGLiteEngine;

@@ -47,6 +28,7 @@ beforeAll(async () => {
 });

 afterAll(async () => {
+  resetGateway(); // don't leak this file's gateway config into shard siblings
  await engine.disconnect();
 });

@@ -3,10 +3,8 @@
 * cycles. Closes codex round-1 P0-5 (write site for last_full_cycle_at
 * was unspecified pre-PR).
 *
- * Conditions for write (keyed off `cycleSourceId` = opts.sourceId ?? the
- * source resolved from brainDir, so the autopilot's inline cycle — brainDir
- * set, no explicit sourceId — also advances the timestamp, #1993):
- *   - a source resolves (explicit sourceId, or brainDir matches a source)
+ * Conditions for write:
+ *   - opts.sourceId is set (legacy callers without sourceId skip the write)
 *   - engine is non-null (no-DB path skips)
 *   - status is 'ok' | 'clean' | 'partial' (failed/skipped don't mark fresh)
 *   - dryRun is false
@@ -92,45 +90,17 @@ describe('runCycle last_full_cycle_at exit hook', () => {
    });
  });

-  test('no explicit sourceId but brainDir resolves a source → writes the resolved source timestamp', async () => {
+  test('legacy caller (no sourceId) does NOT write any source timestamp', async () => {
    await withEnv({ GBRAIN_HOME: gbrainHome }, async () => {
-      // The autopilot's inline cycle sets brainDir but passes no sourceId.
-      // runCycle resolves the source from brainDir (local_path match) into
-      // cycleSourceId and stamps last_full_cycle_at for it — otherwise
-      // cycle_freshness reports the brain stale even while the autopilot
-      // cycles every interval (#1993).
-      await seedSource('resolved-from-dir'); // local_path = brainDir
-      expect(await readLastFullCycleAt('resolved-from-dir')).toBeNull();
-
-      const t0 = Date.now();
-      const report = await runCycle(engine, {
-        brainDir,
-        phases: ['lint'],
-      });
-      expect(['ok', 'clean']).toContain(report.status);
-
-      const after = await readLastFullCycleAt('resolved-from-dir');
-      expect(after).not.toBeNull();
-      expect(new Date(after!).getTime()).toBeGreaterThanOrEqual(t0);
-    });
-  });
-
-  test('no sourceId and brainDir matches no source → does not write', async () => {
-    await withEnv({ GBRAIN_HOME: gbrainHome }, async () => {
-      // A source exists but its local_path does NOT match brainDir, so
-      // resolveSourceForDir returns undefined, cycleSourceId is undefined,
-      // and no per-source timestamp is written.
-      await engine.executeRaw(
-        `INSERT INTO sources (id, name, local_path, config, archived, created_at)
-         VALUES ('unmatched', 'unmatched', '/no/such/repo', '{}'::jsonb, false, NOW())
-         ON CONFLICT (id) DO UPDATE SET local_path = EXCLUDED.local_path`,
-        [],
-      );
+      await seedSource('default-like');
+      // No sourceId passed; should remain untouched.
      await runCycle(engine, {
        brainDir,
        phases: ['lint'],
      });
-      expect(await readLastFullCycleAt('unmatched')).toBeNull();
+      // No per-source write happens; default source's config stays empty.
+      const after = await readLastFullCycleAt('default-like');
+      expect(after).toBeNull();
    });
  });

@@ -0,0 +1,57 @@
+/**
+ * eval-takes-quality gateway self-config — adapter-boundary regression
+ * (takeover of PR #2430).
+ *
+ * The old callsite spread `{ ...cfg, ...process.env }` straight into
+ * configureGateway. The gateway NEVER reads process.env at call time — it
+ * reads `_config.env` — and that spread never populated an `env` field at
+ * all, so every availability/diagnose check dereferenced `undefined.env[k]`
+ * and file-plane API keys (config.json `openai_api_key` etc.) were dropped.
+ * Routing through buildGatewayConfig fixes both. This test fails (throws)
+ * on the old code path.
+ */
+import { afterAll, describe, expect, test } from 'bun:test';
+import { mkdirSync, mkdtempSync, rmSync, writeFileSync } from 'node:fs';
+import { tmpdir } from 'node:os';
+import { join } from 'node:path';
+import { runEvalTakesQuality } from '../src/commands/eval-takes-quality.ts';
+import { isAvailable, resetGateway } from '../src/core/ai/gateway.ts';
+import { withEnv } from './helpers/with-env.ts';
+import type { BrainEngine } from '../src/core/engine.ts';
+
+afterAll(() => {
+  resetGateway();
+});
+
+describe('runEvalTakesQuality — gateway self-config routes through buildGatewayConfig', () => {
+  test('file-plane openai_api_key reaches the gateway env (help path, engine untouched)', async () => {
+    const home = mkdtempSync(join(tmpdir(), 'gbrain-etq-gw-'));
+    try {
+      mkdirSync(join(home, '.gbrain'), { recursive: true });
+      writeFileSync(
+        join(home, '.gbrain', 'config.json'),
+        JSON.stringify({
+          engine: 'pglite',
+          database_path: join(home, '.gbrain', 'brain'),
+          openai_api_key: 'sk-file-plane-test',
+        }),
+      );
+      await withEnv(
+        {
+          GBRAIN_HOME: home,
+          OPENAI_API_KEY: undefined,
+          DATABASE_URL: undefined,
+          GBRAIN_DATABASE_URL: undefined,
+        },
+        async () => {
+          // 'help' returns before touching the engine, but the gateway is
+          // configured first — exactly the seam under test.
+          await runEvalTakesQuality({} as BrainEngine, ['--help']);
+          expect(isAvailable('embedding', 'openai:text-embedding-3-small')).toBe(true);
+        },
+      );
+    } finally {
+      rmSync(home, { recursive: true, force: true });
+    }
+  });
+});