Commit Graph
9 Commits
Author SHA1 Message Date
Story Crater Bot cb52356d13 fix(agent-pod): sync coordinator.js (slug repoId), tighten compaction, route judge to reasoning model 2026-08-19 10:41:33 -07:00
Story Crater Bot d0cbfac7a4 fix(agent-pod): stateless role pool (/new per reuse), never commit PLAN.md 2026-08-19 07:59:01 -07:00
Story Crater Bot 930374a3b8 feat(agent-pod): persistent per-role agent pool, concurrency moves to repo level
coordinator.js now runs one long-lived planner/investigator/implementer/judge
session per repo (reused across every task via tmux send-keys) instead of a
fresh spawn per task per stage. Tasks within a repo run sequentially against
that pool; concurrency is now REPO_CONCURRENCY (default 3) concurrent repos
via a new --repos flag, not concurrent tasks in one repo's phase.
2026-08-18 21:30:49 -07:00
Story Crater Bot 05ff12c117 fix(agent-pod): use process.exitCode not process.exit() in coordinator.js
process.exit() right after console.log() can drop buffered stdout when it's piped (not a TTY) -- exactly kubectl exec's case. Explains the silent empty-output-exit-1 failures. process.exitCode + natural exit lets the event loop drain and flush first.
2026-08-18 19:13:07 -07:00
Story Crater Bot 56b1c96fcf fix(agent-pod): clone deterministically, not through a headless LLM call
git clone is mechanical -- routing it through spawnPi meant a crash gave zero diagnostic output, just a silent exit code. Direct runGit call now, same as commitPending/the squash-merge sequence. Drops the now-unused runStageWithResolver.
2026-08-18 18:57:18 -07:00
Story Crater Bot 26a959215e fix(agent-pod): sync coordinator.js ConfigMap, was stale since auto-discovery landed
The pod's coordinator-src ConfigMap still had the pre-auto-discovery version -- --tasks was required, no task-board parsing, no self-chained stages, no judge model routing. Regenerated from the current source.
2026-08-18 18:42:45 -07:00
Story Crater Bot 193b040de6 feat(agent-pod): implementer and judge learn Playwright for UI verification
Both skills already have Bash in allowed-tools -- no new pi capability needed. For UI/frontend work, implementer screenshots/clicks through the golden path via npx playwright instead of trusting that code compiling means it renders correctly; judge does the same as review evidence, FAILing on visual defects a diff alone wouldn't show. Doesn't apply to non-UI work.
2026-08-18 18:37:46 -07:00
Story Crater Bot 0a87302e19 fix(api): retire Kong key-auth on model routes; agent-pod builds agent-manager fork + ships coordinator.js
Kong key-auth rejected the Authorization: Bearer header every OpenAI-SDK-compatible client sends (verified: raw apikey header works, Bearer doesn't), so it's commented out and stripped from every llm-routes.yaml annotation until there's a Bearer-compatible fix. agent-pod now clones and builds the agent-manager fork from source at container start (no prebuilt binary shipped -- wrong arch and over ConfigMap's size cap) and ships coordinator.js alongside hub.js, so multiple repos can run the pipeline concurrently in one pod via kubectl exec. hub.js keeps its existing role as the container's foreground process, unchanged.
2026-08-18 17:50:52 -07:00
Story Crater Bot c64b68a36b fix(agent-pod): remote tui session for multi-agent 2026-08-18 15:08:04 -07:00