AgentHub/CLAUDE.md

# Claude Code Working Guide

_Last synced: v1.6.0 (2026-09-14) + AI Router profiles (`server/lib/ai-routers.js`, `server/lib/router-apply.js`, the Settings tab `/settings?tool=ai-router`, its **Test** button / `POST /api/ai-routers/test`, the per-lane router select — labelled **Model**, `LaneRouterRow` exported from `LaneCard` and rendered in the Workspace lane-detail header row — and `lanes.router_profile`) + the bootstrap's superseded-build guard and plugin-cache pruner (`supersededBy`/`pruneRetiredPluginRoots` in `scripts/plugin-bootstrap.js`, `unitRoot` in `scripts/systemd-service.js`) + v1.5.6 (2026-09-11) + the systemd handover rule below (`scripts/systemd-service.js`'s `writeUserService`/`restartUserService` split and `plugin-bootstrap.js`'s deferred `pendingSystemdRestart`) + eb63569 (2026-09-10) + release-based update detection (`server/lib/update-check.js` reads the forge releases API — a checkout via its `origin` remote, a plugin-cache install via the marketplace manifest on disk, since an archive-sourced cache dir has no `.git`; a `plugin_managed` install self-updates via the detached `plugin-update` chain) + the install-wide **Update all** batch and session reload (`server/lib/cc-update-all.js`, `server/lib/cc-reload.js`)._ A change that makes any statement in this file, `AGENTS.md`, or a README wrong — a moved path, a new or changed extension point, a retired rule — updates that file in the same change-set and bumps this marker.

## Project mission
- Maintain a reliable local-first dashboard for Claude Code session monitoring.
- Preserve real-time behavior (hooks -> API -> SQLite -> WebSocket -> UI).

## Repo map
- `server/`: Express API, hook ingestion, SQLite access, websocket broadcast (includes release-based update detection — `lib/update-check.js` reads the repo's own releases API and `routes/updates.js` exposes status/check/apply, refusing to self-update a `~/.claude/plugins/cache/` install, plus `lib/workflow-ingest.js` which ingests on-disk Workflow-tool run journals — fleets that emit no hooks — and `lib/codex-ingest.js`, the same idea for `codex` lanes: it tails Codex's own rollout journals because Codex emits no hooks at all).
- `client/`: React + Vite UI.
- `scripts/`: hook installer/handler, import, seed, cleanup utilities. (Update detection lives server-side in `server/lib/update-check.js`, keyed on the forge's **newest release** rather than a commit count. `POST /api/updates/apply` has two modes and picks by install kind, never from the request: a fast-forwardable **checkout** merges the release tag `--ff-only`, rebuilds, and self-restarts via `server/lib/self-restart.js` + the detached `scripts/restart-helper.js`; a **plugin-cache** install is handed to `server/lib/plugin-update.js` + the detached `scripts/plugin-update-helper.js`, which runs `claude plugin marketplace update` → `claude plugin update -y` → `plugin-bootstrap.js --force` from the NEW cache dir. Both write the HTTP response before anything stops the process, and every plugin-chain outcome is recorded to `plugin-update.json` in the runtime dir — outside the cache, so it survives the swap and the next server reports it as `last_plugin_update`. Neither mode can run `/reload-plugins`: it is REPL-scoped to the session that invokes it, so an open Claude Code session still reloads by hand for new skills/commands. Elsewhere users run the printed manual command, surfaced in the UI and by `agenthub update-check`. `scripts/cc-update-all-helper.js` is the same detached pattern for the config panel's **Update all** button — marketplaces → every plugin → AgentHub LAST → one reload — journalling each step to `cc-update-all.json` in the runtime dir.)
- `skills/`: this repo's skills. Because the repo root **is** the `agenthub` plugin root, this is also the plugin's skills directory — every skill here ships to anyone who installs the plugin, addressed as `/agenthub:<name>` (bare `/<name>` works as an alias). Put a skill here only if it should reach plugin users; `agenthub skills install` remains a fallback that copies `ship-feature-lane` into `~/.claude/skills/` for plain checkouts. - `agents/`: the same story for subagents — plugin root, so all four (`qc-local`, `senior-gate-reviewer`, `backend-reviewer`, `frontend-reviewer`) ship as `agenthub:<name>`. This is also the source `agenthub lanes agents install` copies the two pipeline agents from into a lane's `.claude/agents/`, which is what keeps the bare `subagent_type: qc-local` working there. One file per agent, never a second copy.
- `.claude-plugin/`: the marketplace plus the root `agenthub` plugin manifest (`"source": "./"` — the whole repo is the plugin). Its hooks are inline in `plugin.json`; its commands live in `plugins/agenthub/commands/`, which is NOT a subdirectory plugin. `scripts/plugin-bootstrap.js` runs from `SessionStart` and owns the writable runtime under `~/.claude/agenthub/runtime/` — it never writes into the plugin cache, which Claude Code replaces on every update. See `docs/PLUGINS.md`.
- **Reporting a required plugin is derived; installing one is allowlisted.** `server/lib/skill-plugins.js` reads the `/plugin:skill` strings out of the loaded pipeline templates — right for the `/agenthub-doctor` row, wrong for an install: a template is a file anyone can drop into `DASHBOARD_PIPELINES_DIR`, so installing whatever name it contains would make "edit a JSON file" mean "run this vendor's install command next session". The bootstrap installs only what `AUTO_INSTALLABLE` (a constant in that file, `superpowers` today) lists, in its DETACHED worker, never in the 10 s hook, recorded once as `state.skillPluginsSettled` (also gating `fastPathOk`), never throwing, and off under `AGENTHUB_AUTO_INSTALL_PLUGINS=0|false|off`. Do not derive an install target from a template.

## Lanes

A **lane** is a durable unit of parallel agent work — one working directory, many Claude Code sessions over time. Lanes are keyed by `cwd`, not `session_id`, so they survive session restarts and allow the dashboard to show a pipeline that persists across runs.

**Key points:**
- Sessions are bound to lanes by longest path-boundary prefix match on their `cwd` (set in hook data from `claude` CLI).
- A lane's stage is **declared** by the driving Claude session (via `agenthub stage <stage> [--evidence "..."]`), not inferred by the dashboard.
- **A CLI command resolves its lane by (cwd, tool), never cwd alone.** `resolveLaneArg` (`bin/agenthub.js`) answers in order: `--lane`/leading id → `AGENTHUB_LANE_ID` (every tmux pane the dashboard starts carries it, via `tmux new-session -e`) → `--tool`, else the harness marker (`CLAUDECODE` ⇒ claude; Codex has NO reliable env marker, so it passes `--tool codex`) → the owning lane. Two candidates at one directory and no signal is an ERROR that names both, never a pick — writing a codex session's stage onto the claude lane's card is a lie about which agent did the work.
- **The console never writes a lane's stage.** Declared stages come from `agenthub stage` (a skill calling the CLI command), and inferred stages come from tool-event detection only. **One exception, `codex` lanes only:** `codex-ingest.js` advances `stage` from its own written detection (`setStage(..., { keepDetection: true })`) — a codex session has no hook and rarely calls `agenthub stage`, so nothing else would ever move its map. It passes NO evidence, so the node renders `passed-no-evidence` (amber) and never `done`, and `keepDetection` preserves the detection columns so the card still shows the command that moved it. Do not extend this to `claude` lanes or to the console. The UI's run console and the Workspace page never call `POST /api/lanes/:id/stage` — that boundary exists because the console observes tool calls but cannot see their outcomes. A lane's stage is the session's truth about progress; the console's observation of tool execution would not be proof of completion.
- **Going back is a declaration, and it takes the nodes ahead with it.** A declaration landing on a node BEFORE the lane's current one (a fix-loop re-entry, or a new request on a lane that already reached `done`) clears the detection columns like every real transition AND drops the `stages` records of every node ahead of it — those belong to the previous pass, and left in place they render `done` with the old work's evidence beside a `current` ring further back, contradicting `progress` (computed from the declared index alone). Records under a key matching no node are left alone: unplaceable, so unprovably ahead. `agenthub feature activate <slug>` is the other half of this — a feature switch archives the outgoing pipeline and clears detection too, because a detection belongs to the feature that produced it.
- **A detection expires, an evidence rule does not.** `recordDetection` skips its forward-only comparison once `detected_at` is older than `DETECTION_TTL_MS` (default 5 min), so a lane can move backwards between work sessions. That window changes only WHICH detection is current — it never relaxes declared-wins (an agent's own claim has no expiry) and never lets an inferred node render `done`. Declared-wins has exactly ONE exception: a declaration on the pipeline's LAST node (`done`) stops vetoing, because the pipeline is over there and a later detection is new work — without it a lane left at `done` rejects every detection for the rest of its life.
- **Working-copy facts live at `GET /api/lanes/:id/git`, never inside `GET /api/lanes`.** That endpoint shells out to git three times; the lane list is polled and re-broadcast on every hook. A cwd that is not a readable repo returns `{available:false}` with HTTP 200 — a normal state, not a fault. Cards fetch it themselves every 30s and fail silently.
- **The Workspace console collapses with CSS, never by unmounting — and it has exactly ONE render site.** Unmounting `RunConsole` disposes the run subscription and drops a live run's rendered history. The same rule binds one level up: `ConsoleArea` lives inside the always-rendered `lane-detail` section in `client/src/pages/Workspace.tsx`, never inside a `{currentLane && …}` branch. A second render site makes React unmount the whole console every time `currentLane` flips — lanes arriving on mount, the selected lane removed, a refetch re-picking the selection — which kills the terminal and lets `LaneConsolePane`'s lane-switch effect wipe a half-typed prompt. Header and info block are the conditional parts; the console is not.
- **AgentHub does not orchestrate:** no chaining, no queue, no retry logic, no gate evaluation. The session in control makes all decisions; the dashboard records the claimed stage and shows evidence. The runtime layer (`agenthub lanes up|down|hook`) does not change this — it offers *primitives* a session calls; nothing in the dashboard sequences them.
- **The runtime never writes `stage`, `status` or `notes`** — only `slot` and `ports`. `status=running` means an AGENT is working, not that a server is listening; merging the two would corrupt `classifyLiveness`. Boot failures live in `LANES_ROOT/.state/lane<slot>/last-error.json` and surface via `GET /api/lanes/:id/runtime`. Same boundary as the console-never-writes-stage rule above.
- **A lane's stack is up or down as a computed fact, never a stored one.** `runtimeFacts` re-derives it from pid files and port probes on every read. A process dies to OOM, a stray `kill`, a reboot — caching a truth AgentHub does not control buys ghost state. Slot allocation is the opposite (fully controlled, must be race-free), so that one does live in the DB, under `withLaneLock` plus a partial unique index.
- **A profile's `profile.env` is parsed, never sourced.** Hooks are executed deliberately; config is only read. Sourcing arbitrary shell from a user's repository into the dashboard process would be a code-execution path. Hook names always come from the fixed allowlist in `server/lib/lane-profile.js`, never from a request.
- Pipeline templates are JSON files (`server/data/pipelines/`) with node definitions; custom templates override built-ins when `DASHBOARD_PIPELINES_DIR` is set.
- Nodes render in five states: `failed` (rejected), `current` (now), `done` (with evidence), `passed-no-evidence` (claimed or skipped, amber), `pending` (not reached).
- Liveness: a silent **watcher** (stage matching `/watch|poll/`) is dead after `LANE_DEAD_SEC` seconds (default 300); a silent **idle** lane is at rest, not dead.
- **Never `rm -rf` a lane.** Destructive lane operations (`reset`, `remove`, `purge`) go through `server/lib/worktree.js`/`server/lib/lanes.js`, never a raw filesystem delete — that's what keeps the three-check destroy guard and the DB bookkeeping in lockstep.
- **Never build a git command as a shell string.** Lane git operations use `execFile("git", [...args])` with an explicit argument array (see `server/lib/worktree.js`'s `git()` helper), never a concatenated/interpolated string passed to a shell.
- **A lane's AI Router profile is a lane fact too, and it never travels alone.** `lanes.router_profile` (nullable; null means the same as `native`) is written ONLY by `setRouterProfile`, called beside the disk write in `POST /api/lanes/:id/router/apply` — never through `PATCHABLE`. The column and the lane's own config must move together, or the row claims a router the lane's `settings.local.json` has never heard of. What is written comes from the lane row, never the request: a **claude** lane gets eight `env` keys merged into `<cwd>/.claude/settings.local.json` (`.local` because the value carries an API key; every other key in that file is preserved and a non-JSON file is refused, not rewritten), a **codex** lane gets `$CODEX_HOME/<name>.config.toml` — the layer `codex -p <name>` loads — because Codex has NO project-level config, and its base `config.toml` is never touched. Key material never leaves the server: profiles live at `<CLAUDE_HOME>/agenthub-routers/` (0700/0600), `GET /api/ai-routers` reports only `hasKey`, and there is deliberately no route that returns one profile in full. **`POST /api/ai-routers/test` sends a stored key ONLY to the endpoint it was stored against** — a blank `apiKey` falls back to the profile named by `name`, but only while the submitted `baseUrl` still equals the saved one, else `EBADKEY`. Relax that and the route becomes a way to make the server POST a saved key to any URL a browser names. **Re-pointing a lane is allowed and is not a PATCH:** the lane card's AI Router select calls `POST /api/lanes/:id/router/apply` (column + disk together, same rule as above), never `PATCH /api/lanes/:id`. **Codex's key reaches the pane through tmux's `-e` argv and is never typed into it** — `tmux.sendCommand` takes over an idle pane with `respawn-pane` for exactly this reason; a `KEY='sk-…' codex …` line typed with `send-keys` would be rendered to every attached browser, kept in the scrollback and appended to the user's shell history. Do not reintroduce a typed command line there. **And attaching COPIES a profile**, so the store stops being the single source of truth the moment a lane exists: `PATCH`/`DELETE /api/ai-routers/:name` must push through to the lanes naming it (`reapplyToLanes`/`detachFromLanes`), or a rotated key silently never reaches the lane still using the revoked one. See `server/lib/ai-routers.js`, `server/lib/router-apply.js`.
- **A lane's `tool` is a lane fact, immutable, and never read from a request.** `lanes.tool` (`claude` | `codex`, default `claude`) is set by `createLane` only — it is deliberately absent from `PATCHABLE`, so no PATCH can flip the CLI under a live pane. Every spawn path (`POST /api/lanes/:id/start`, `POST /api/run`) takes the tool from the lane row and ignores whatever the body says; the run panel's Codex-only fields (`sandbox`, `approval`, `resumeLast`) are validated against the fixed allowlists in `server/lib/pty-run.js` before they reach an argv, and the bypass flags (`--dangerously-bypass-approvals-and-sandbox`, `--dangerously-bypass-hook-trust`) are never emitted. `GET /api/run/binary?tool=` resolves the probed binary name from that same allowlist — the query string reaches `which`.
- **Adopted lanes are not destroyable.** A lane with `kind === "adopted"` may never have its worktree reset or removed; "removing" one only drops the dashboard's own record of it. This is enforced in code (`assertDestroyable`'s first check) — do not add a path that bypasses it.
- **A `cwd` can back at most one lane per tool, not one lane overall.** Uniqueness is the composite `idx_lanes_cwd_tool` (`cwd`, `tool`), not `cwd` alone — Repo/Adopt mode (`POST /api/lanes/ensure`) lets the same directory carry a `claude` lane and a `codex` lane side by side, shown as two separate cards. `resolveLaneByCwd(cwd, tool)` takes an optional `tool` filter for exactly this; hook ingestion (`touchLaneFromHook`, `resolveLaneForWork`) always pins it to `"claude"` because Codex has no hook mechanism in this codebase, so a hook event must never touch a `codex` lane sharing the path. Worktree mode is unaffected — it already allocates a unique directory per lane.

See `docs/LANES.md` for full guide: stage reporting, custom templates, lane actions, and the five node states.

## Non-negotiable engineering rules
- **When a `agenthub.service` unit of ours owns the server, nothing else may stop it.** The bootstrap that performs a self-update is spawned by the running server, so it lives in that server's systemd cgroup: SIGTERM-ing the server from there exits 0 (which `Restart=on-failure` read as "meant to stop") AND trips `KillMode=control-group`, reaping the bootstrap before it can repoint the unit or write `state.json` — a dashboard dead after every update until some later session bootstrapped again. So `scripts/systemd-service.js` splits the API: `writeUserService` repoints the unit and starts NOTHING, `restartUserService` uses `--no-block`, and `plugin-bootstrap.js` defers that restart past `writeState` and `releaseLock` (`pendingSystemdRestart`). The unit is `Restart=always`. Do not add a path that stops the dashboard from inside its own cgroup, and do not move the restart earlier. `agenthub service install` may keep using `installUserService` — it runs from a user's shell, outside the cgroup. `systemdOwned` additionally requires the unit to be ACTIVE (an installed-but-stopped unit owns nothing; treating it as owner skips the stop and then binds an occupied port). A unit naming a RETIRED plugin root is invisible until a reboot, so `fastPathOk` rejects it via `unitIsStale` (pure fs — that path may never spawn a process) and the bootstrap rewrites it; repairing the unit is NOT restarting the server (`pendingSystemdRestart = systemdOwned && needsServer`), and the detached-spawn fallback is guarded by `needsServer` so a stale unit alone never starts a second server onto a bound port.
- **A bootstrap that is no longer the installed build does NOTHING, and `--force` does not override that.** `supersededBy` (`scripts/plugin-bootstrap.js`) reads Claude Code's `installed_plugins.json`: if a SIBLING cache directory is the recorded install, this build logs one line and returns `superseded` — no deps, no unit, no server. A Claude Code process that was not restarted after an update keeps loading the OLD cache dir, and without this its every session repointed `agenthub.service` at that old dir and restarted the dashboard onto the version the user just upgraded away from, once per session, forever. It may not hand over to the new build either — the process it would replace is the one loading it. Do not add a path that lets a retired root write the unit, stop the server, or "just" refresh deps.
- **Deleting retired cache copies is the ONE thing here that touches the plugin cache, and four protections gate it.** `pruneRetiredPluginRoots` never deletes: this build; any `installPath` in `installed_plugins.json`; the dir `agenthub.service` names (`unitRoot`, foreign unit included); any dir named in the user's `settings.json` hooks (a un-restarted Claude Code is still executing from there — deleting it breaks that session's hooks). Plus the 2 newest survivors as a rollback cushion, and nothing at all outside `~/.claude/plugins/cache/`. It runs AFTER `writeState`, so a bad prune cannot also cost the record of a good bootstrap. Adding a fifth thing that pins a cache dir means adding a fifth protection here.
- **A session reload is typed into a live pane, never spawned.** `/reload-plugins` and `/reload-skills` are scoped BY CLAUDE CODE to the session that runs them, so nothing can reload a session from the outside. `server/lib/cc-reload.js` therefore types them into the dashboard's existing tmux lane panes and spawns NO `claude` process — a one-shot `claude -p "/reload-plugins"` reloads only its own short-lived session, exits, and leaves a junk session in the dashboard's own history. Do not add a path that starts a session in order to reload one.
- Preserve existing behavior unless explicitly asked to change it.
- Prefer minimal, reversible diffs.
- Never silently weaken safety controls around destructive actions.
- Keep docs updated when behavior, commands, file locations, or workflows change — apply the `update-project-docs` skill automatically at the end of every change-set that alters behavior, config, interfaces, events, schema, CLI commands, or features (do not wait to be asked).
- Every applicable source file you create or update (`.js/.ts/.tsx/.cjs/.mjs/.py/.sh/.css`) must start with the copyright/authorship header — file overview + the exact line `@author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>`. Formats and audit script: `skills/file-headers/` (verify with `bash skills/file-headers/scripts/check-headers.sh`). This binds every coding agent (Claude Code, Codex, or others).

## Commands you should know
- Setup: `npm run setup`
- Dev: `npm run dev`
- Prod build/start: `npm run build` then `npm start`
- Server tests: `npm run test:server`
- Client tests: `npm run test:client`
- CLI (after setup): `agenthub <command>` — terminal access to the full dashboard surface (`bin/agenthub.js`; `agenthub help` lists commands)

## Testing and verification policy
- Backend changes: run `npm run test:server` before finishing.
- Frontend changes: run `npm run test:client` when relevant. This includes per-screen render snapshots (`client/src/pages/__tests__/screens.snapshot.test.tsx`). If a UI change is intentional, review the snapshot diff and regenerate baselines with `cd client && npx vitest run -u`; never blindly update snapshots to make tests pass.
- If you cannot run a verification step, state exactly what was not run and why.

## Change guidelines by area
- API routes: preserve response shapes unless change is requested and documented.
- Database: avoid schema changes without migration-safe logic.
- Hooks: keep fail-safe and non-blocking behavior.
- WebSocket: keep message types stable and backward-compatible.
- Documentation: include exact commands and paths; keep markdown examples runnable.

## Agent behavior
- Explore first, then implement.
- For larger tasks, propose/check a short plan before broad edits.
- Use file-specific rules in `.claude/rules/` when working in scoped areas.
- Use project skills from `skills/` for repeatable workflows (they ship with the `agenthub` plugin: `/agenthub:<name>`, or the bare `/<name>` alias).
- Use `agents/` subagents for focused review or investigation passes (they ship with the plugin: `agenthub:backend-reviewer`, etc.).
- **Declare lane stage even in plain chat, not just inside skills.** `agenthub stage` (see `docs/LANES.md` § Reporting a stage) is a reporting command, not a skill-only ritual — any Claude session working inside an adopted lane's cwd should call it on real stage transitions (starting to plan, starting to implement, running tests, opening the PR, etc.), whether or not a skill is driving. Tool-event detection (`server/lib/stage-detect.js`) only ever paints the amber "detected" badge, never the blue `current` ring — a lane worked entirely through plain chat with no `agenthub stage` calls will show a stale `current` stage no matter how much real work happens. Skip it only when `agenthub stage` reports no lane owns the cwd (not adopted).
- **Prefer `agenthub:ship-feature`(`-lane`) over invoking `superpowers:brainstorming`/`writing-plans`/`executing-plans`/`subagent-driven-development`/`test-driven-development` directly for implementation work inside a lane's cwd.** Those Superpowers skills know nothing about AgentHub lanes and never call `agenthub stage` themselves — going straight to them (skipping the AgentHub wrapper) is exactly how a lane's `current` stage goes stale mid-implementation while `agenthub stage` is never called for real transitions. `agenthub:ship-feature`/`ship-feature-lane` drive the same underlying Superpowers skills phase-by-phase and declare `agenthub stage` at every boundary — pick one of them first for any multi-phase feature/bugfix inside a lane; reach for a bare Superpowers skill directly only for work with no lane cwd to report against (e.g. a scratch repo, or explicitly told to skip the AgentHub wrapper).