* Initial plan
* chore: bump all workspace package versions from 0.2.5 to 0.3.0
Agent-Logs-Url: https://github.com/ComposioHQ/agent-orchestrator/sessions/da5b2769-e7d4-4d08-a60c-bd5f695d1ca7
Co-authored-by: harshitsinghbhandari <212377671+harshitsinghbhandari@users.noreply.github.com>
* fix: update package-version test to expect 0.3.0
Agent-Logs-Url: https://github.com/ComposioHQ/agent-orchestrator/sessions/ad61e33e-417f-4482-b06c-0b60826b7f2d
Co-authored-by: harshitsinghbhandari <212377671+harshitsinghbhandari@users.noreply.github.com>
* chore: revert non-ao version bumps
Only @aoagents/ao drives the 'ao update available' prompt
(packages/cli/src/lib/update-check.ts compares against the
@aoagents/ao registry version and reads the local @aoagents/ao
package.json). All other workspace bumps are unnecessary.
* chore: align workspace versions with npm registry
Catch up source-of-truth package.json versions to what is already
published on npm. The registry reflects releases done via Changesets;
the in-tree files had drifted to 0.2.5.
0.2.5 -> 0.3.0: cli, core, web, agent-aider, agent-claude-code,
agent-codex, agent-opencode, notifier-composio,
notifier-desktop, notifier-slack, notifier-webhook,
runtime-process, runtime-tmux, scm-github,
terminal-iterm2, terminal-web, tracker-github,
tracker-linear, workspace-clone, workspace-worktree
0.2.5 -> 0.2.6: notifier-discord, notifier-openclaw, scm-gitlab,
tracker-gitlab
0.1.0 -> 0.1.1: agent-cursor
Also updates agent-codex package-version.test.ts to expect 0.3.0.
* test(cli): use future version in update-check cache test
The cache-fresh test assumed getCurrentVersion() returned a value
older than the cached latestVersion. With packages/ao now at 0.3.0
and resolvable from cli via pnpm's hoisted store at test time,
getCurrentVersion() returns 0.3.0, so isOutdated against a cached
latestVersion of 0.3.0 is false and the assertion fails.
Use 99.0.0 in the cache so the comparison stays meaningful regardless
of the current installed version.
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: harshitsinghbhandari <212377671+harshitsinghbhandari@users.noreply.github.com>
Co-authored-by: harshitsinghbhandari <claudeagain@pkarnal.com>
* fix: reduce opencode session list churn
* fix: make bun-tmp-janitor cross-platform and move to process-level boot
- Extend platform support from Linux-only to Linux + macOS (win32 skipped
since opencode ships no Windows binary and the kernel disallows unlinking
mapped files there)
- Use os.tmpdir() instead of hardcoded /tmp to handle macOS $TMPDIR paths
- Extend file pattern from \.so to \.(so|dylib) to cover macOS dylib leaks
- Move startBunTmpJanitor() from ensureLifecycleWorker() (per-project) to
the process-level boot in registerStart() immediately after register(),
where the single-instance contract is already in force
- Drop the project-observer-bound onSweep closure that incorrectly attributed
janitor health to whichever project happened to start first; replaced with
a simple stderr warn on errors (no project context needed for a process-wide
sweep of /tmp)
- Move stopBunTmpJanitor() into the SIGINT/SIGTERM shutdown handler in
registerStart() alongside stopAllLifecycleWorkers()
- Remove startBunTmpJanitor/stopBunTmpJanitor from lifecycle-service.ts
entirely; lifecycle workers have no business knowing about a process-wide
OS resource
* fix(opencode): address PR #1478 review (TMPDIR isolation, shared cache, janitor cleanup)
Implements all seven findings from the PR #1478 review:
Core / agent-opencode:
- New @aoagents/ao-core/opencode-shared module owns the single TTL cache
+ in-flight dedup for 'opencode session list' (was duplicated across
core and the plugin, doubling spawns per poll cycle).
- TTL dropped from 3s to 500ms so the send-confirmation loop's
updatedAt > baselineUpdatedAt delivery signal can actually fire.
- New invalidateOpenCodeSessionListCache() called by deleteOpenCodeSession
so reuse / remap / restore code paths cannot observe a deleted id.
- New getOpenCodeChildEnv() / getOpenCodeTmpDir(): every opencode child
spawned by core, the plugin, or the agent runtime points TMPDIR/TMP/TEMP
at ~/.agent-orchestrator/.bun-tmp. Bounds the janitor's blast radius
to AO-owned files even on shared hosts.
CLI janitor:
- Sweeps only the AO-owned tmp dir (not the system /tmp).
- Filters synchronously before spawning per-entry stat/unlink work.
- stopBunTmpJanitor() is now async and awaits any in-flight sweep so
SIGTERM cannot exit while unlink() is mid-flight; start.ts shutdown
handler awaits it.
- onSweep callback in start.ts now logs successful reclaims, not just
errors, so operators can confirm the janitor is doing useful work.
Tests:
- packages/core/__tests__/opencode-shared.test.ts (TTL contract,
TMPDIR location, env merge semantics).
- packages/cli/__tests__/lib/bun-tmp-janitor.test.ts (sweep behavior,
stop-awaits-in-flight, pattern matching, missing-dir tolerance).
* chore: remove review postmortem artifact
* fix(cli): remove start command non-null assertions
* feat(core): add opt-in gh CLI tracer and migrate scm/tracker plugins
Introduces execGhObserved() in @aoagents/ao-core: a thin wrapper around
execFile("gh", ...) that writes a JSONL trace row to $AO_GH_TRACE_FILE
on both success and failure. Captures status line, HTTP status, ETag,
rate-limit headers, duration, stdout/stderr byte counts, exit code, and
signal. No-op when the env var is unset, so default behavior is
unchanged.
Migrates three call sites to the observer:
- scm-github/graphql-batch.ts — PR-list guard, commit-status guard,
GraphQL batch query
- scm-github/index.ts — gh() and ghInDir() helpers
- tracker-github/index.ts — internal gh() helper
This is Phase A1a of experiments/PLAN.md: tracer infrastructure +
migration. The full GhRunner contract (Promise<GhResult>,
GhRunnerError.ghResult on reject, body capture, redaction, 64 KB cap)
lands in A1b along with the scorecard baseline.
Also adds experiments/ reference docs: the v2.3 plan, the gh-CLI call
catalog, two ETag verification writeups, and a trace harness + summary
script.
* docs(experiments): add A1a validation status and A1b blockers
Record the five A1b pre-freeze blockers surfaced by Adil's 1,487-row
baseline and an independent drill run: graphql-batch missing -i,
extractOperation flag mis-bucketing, analyzer not segmenting burn by
reset window, CLI-subcommand opacity (GH_DEBUG=api stderr vs coarse
/rate_limit bracket — not equivalent), and sessionId/projectId not
threaded through plugin callsites. Note bare gh() helper cleanup as
known-open follow-up.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(tracer): close A1b blockers 1-4 — graphql-batch visibility, operation naming, analyzer segmentation
- Add -i flag to executeBatchQuery in graphql-batch.ts and split HTTP
headers from JSON body before parsing, making all gh.api.graphql-batch
rows visible to status and rate-limit analysis (was 186 invisible rows)
- Fix extractOperation() in gh-trace.ts to walk past -* flags before
picking the operation segment, eliminating the gh.api.--method bucket
- Add per-reset-window burn segmentation to both analyzers so runs
straddling a reset boundary produce per-window deltas instead of a
single invalid cross-reset delta
- Add experiment scripts: analyze-trace.mjs (deep trace analysis) and
drill-tracer.mjs (standalone tracer exerciser)
- Document Gap 1 decision in PLAN.md: accept CLI subcommands as opaque
for A1, bracket A2 runs with /rate_limit snapshots for coarse burn
- Add progress timeline to PLAN.md showing A→B→C track dependencies
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* docs(experiments): add A2 baseline matrix runbook
Practical execution plan for the Phase A2 scenario x scale x topology
matrix: 7 priority cells, per-cell procedure, /rate_limit bracketing
for Gap 1 subcommand burn, output format for baseline.md, and the
scorecard that gates Track B.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(tracer): guard stderr/stdout against undefined, bound operation cardinality
Addresses code review findings:
1. Guard Buffer.byteLength and parseIncludedHttpResponse against
undefined stderr/stdout — fixes 48 SCM test regressions where
mocked execFile paths don't populate stderr
2. extractOperation() now takes only the first path segment of REST
URLs (e.g. "repos" from "repos/acme/repo/pulls/123/...") to keep
operation bucket cardinality bounded and stable across runs
3. Fix A2 runbook /rate_limit snapshots to produce valid JSON using
jq's now|todate instead of appending raw timestamp
4. Add blocker 5 dependency to runbook prereqs and per-session cells
All 140 SCM tests pass (0 failures).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* docs(experiments): add rate-limiting research artifacts
Baseline measurements, discussion notes, benchmark harness spec,
and updated master plan from two independent trace runs at 5-6 sessions.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(experiments): add benchmark harness for GH rate-limit measurement
Three modes: setup (spawn sessions, wait for PRs), measure (trace API
calls over a fixed window, produce scorecard), report (recompute from
existing trace). Node.js stdlib only, shells out to ao CLI and gh CLI.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(scm-github): handle 304 Not Modified in ETag guard catch blocks (B1)
`gh api -i` exits code 1 on HTTP 304 responses, causing the catch blocks
in checkPRListETag and checkCommitStatusETag to assume the resource changed
and trigger unnecessary GraphQL batch queries every poll cycle.
Fix: inspect stdout/stderr in the catch block for the 304 status line before
falling back to "assume changed". Also unifies the 304 detection regex to
handle HTTP/1.1, HTTP/2, and HTTP/2.0 status lines, and adds rateLimit
introspection to the batch GraphQL query.
Benchmark result (quiet-steady, 5 sessions, 15 min):
- GraphQL points/hr: 260/5,000 (5%) — down from 820–1,416 pre-fix
- ETag guard 304 rate: 100%
- GraphQL batch calls during measurement: 0
Also fixes the benchmark harness to create placeholder tmux sessions with a
claude symlink so the lifecycle actually polls sessions instead of
short-circuiting to "killed".
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* docs(experiments): update plan and notes with B1 benchmark results
B1 fix validated at 5, 10, and 20 sessions in quiet-steady state:
- 5 sessions: 260 GraphQL pts/hr (5% budget)
- 10 sessions: 640 pts/hr (13%)
- 20 sessions: 680 pts/hr (14%) — sub-linear scaling confirmed
- 50-session projection: ~800-1000 pts/hr (16-20%)
- ETag guard 304 rate: 100% at all scale points
- graphql-batch calls: 0 during measurement at all scale points
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(core): log gh wrapper invocations for D1
* fix(core): preserve wrapper logging for dash-prefixed gh args
* feat(core): add gh wrapper cache for PR discovery and issue context (D4)
Add read-through caching to the ~/.ao/bin/gh wrapper, targeting the two
largest agent-side waste buckets identified in D4 analysis:
1. PR discovery (gh pr list --head): infinite TTL for positive results.
598 calls → ~10 per 10-session run (98% reduction).
2. Issue context (gh issue view): 300s TTL.
75 calls → ~20 per 10-session run (73% reduction).
The wrapper now caches successful read-only responses in
$AO_DATA_DIR/.ghcache/$AO_SESSION/ and serves them on subsequent
identical calls. Negative results (empty []) are never cached.
gh pr create populates the PR discovery cache immediately.
Also lifts PATH wrapper installation from individual agent plugins into
session-manager, making it universal for all agents including Claude Code:
- session-manager injects PATH + GH_PATH into every runtime.create()
- session-manager calls setupPathWrapperWorkspace() for all agents
- Removes duplicate buildAgentPath/setupPathWrapperWorkspace boilerplate
from codex, aider, opencode, and cursor plugins
Includes D4 implementation plans in experiments/.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add full capacity discovery (5→50 sessions) and CI churn results
Complete scaling curve measured: 50 sessions uses only ~28% of GraphQL
budget with 100% ETag guard hit rate at every scale. Poll cycle lag
identified as first bottleneck (66s at 50 sessions vs 30s target).
CI churn benchmark shows ETag invalidation is a latency problem, not
a rate-limit problem (+9% GraphQL, +4.4x p50 latency).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* docs(experiments): record real-agent catastrophe and Track D handoff
5-real-agent run on todo-app exhausted GraphQL bucket in 31 min (~9572 pts/hr,
~37x quiet-steady at the same session count). AO polling consumed ~10 calls;
the rest came from agents themselves via the metadata-only ~/.ao/bin/gh
wrapper, which has no tracing. Captures findings, adds Track D (agent-side
gh consumption) plus B5 (migrate remaining bare gh callsites to
execGhObserved), and includes the runbook + benchmark scripts Adil will
build on for the cross-machine reproduction.
* feat(core): add cache-hit/miss tracing to gh wrapper (D4)
The wrapper trace now logs a cacheResult entry for every cacheable
command: hit, miss-stored, miss-negative, or miss-error. This makes
benchmark runs conclusive — you can count cache hits vs real gh calls
directly from the JSONL trace instead of inferring from rate-limit
deltas.
Bump wrapper version to 0.4.1.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* feat(scm-github): replace repo-scoped Guard 1 with PR-scoped ETag checks (D4)
Guard 1 now checks GET /repos/{owner}/{repo}/pulls/{number} per PR
instead of GET /repos/{owner}/{repo}/pulls?... per repo. This means:
- Only changed PRs flow into the GraphQL batch
- Unchanged PRs are served directly from the enrichment cache
- shouldRefreshPREnrichment returns a refresh plan (prsToRefresh +
cachedResults) instead of a boolean
When 1 of 10 PRs changes, the old guard refreshed all 10 via GraphQL.
Now only the 1 changed PR is fetched; the other 9 are served from cache
at zero GraphQL cost.
Trade-off: more REST guard calls (1 per PR instead of 1 per repo), but
304 responses cost zero rate limit points.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* Forward AO_AGENT_GH_TRACE to session runtimes
* revert: remove PR-scoped ETag guards (Change 3)
Reverts 25ae6013. The per-PR Guard 1 added more REST calls (1 per PR
instead of 1 per repo) without meaningful GraphQL savings at 10-session
scale. Core REST delta went from 16 to 142 while GraphQL rate stayed
flat. The repo-scoped guard is sufficient for current workloads.
Preserves the subsequent 6fc64f4f commit (AO_AGENT_GH_TRACE forwarding).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(core): use real gh binary in execGhObserved, bypass wrapper
execGhObserved() was calling bare "gh" which resolved to ~/.ao/bin/gh
(the wrapper) when that directory was in PATH. This caused:
- AO-side gh calls going through the agent wrapper
- All trace rows with aoSession=null polluting the agent trace
- Cache functions silently failing (no AO_SESSION in AO process)
Now strips ~/.ao/bin from PATH and resolves the real gh binary
(e.g. /opt/homebrew/bin/gh) at startup. Cached after first resolution.
AO process → execGhObserved → real gh → AO_GH_TRACE_FILE
Agent process → ~/.ao/bin/gh wrapper → AO_AGENT_GH_TRACE + cache
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(core): harden gh wrapper caching and agent-side tracing
Cache correctness:
- Include --json fields in cache key (prevents stale partial responses)
- Only cache stdout, not stderr (prevents warning contamination)
- Fix trailing newline inconsistency in PR discovery cache
- Support --key=value arg syntax for all cached flags
- Remove PR create cache pre-population (hardcoded fields, no JSON escaping)
- Log miss-write-failed when ao_cache_write fails (previously silent)
Agent trace improvements:
- Add operation field to invocation rows (gh.pr.list, gh.issue.view, etc.)
- Add durationMs, exitCode, ok to cache outcome rows
- Log passthrough for all non-cached code paths (pr/create, default case)
- Replace exec with child process in default case to enable post-call tracing
Bump wrapper version to 0.6.0.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(runtime-tmux): re-export PATH after shell init to survive macOS path_helper
macOS zsh runs path_helper during shell startup which resets PATH,
wiping entries set via tmux new-session -e. This caused ~/.ao/bin
to be lost, so the gh/git wrappers were never intercepting agent
calls — no caching, no tracing, no metadata auto-updates.
Fix: send `export PATH=...` via send-keys after the shell has
initialized but before the launch command, ensuring PATH sticks.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* fix(runtime-tmux): use launch script for PATH re-export instead of send-keys
The previous send-keys approach sent 1000+ literal keystrokes for the
PATH value, which broke terminal input buffers and caused stuck quote
prompts. Instead, include the PATH export in the launch script file
which is executed directly — no terminal buffer issues.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add AO-side gh rate-limit trace report
5-session, 15-minute trace analysis with full call breakdown,
ETag guard effectiveness, anomaly investigation, and ranked
reduction opportunities.
Key findings:
- GraphQL at 41%/hr with 5 sessions (bottleneck at ~12 sessions)
- 47% of calls are individual REST fallbacks that batch should cover
- Review thread GraphQL calls (55/15min) can be folded into batch
- detectPR() and guard failures are working as designed
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add AO rate-limit reduction plan with Step 1
Step 1: Remove individual REST fallback from determineStatus().
110 calls (65 pr view + 45 pr checks) eliminated per 15-min window.
Batch enrichment covers all PRs every 30s — fallback is unnecessary
insurance for an event that never occurred in real traces.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* experiments(m2): drop agent-trace gate for claude-code
Claude Code uses native PostToolUse hooks (.claude/settings.json), bypassing
the ~/.ao/bin/gh PATH wrapper, so AO_AGENT_GH_TRACE stays empty even when
Claude makes gh calls. The previous smoke gate required AGENT_ROWS>0 and
aborted every claude-code M2 batch at smoke.
- limit-finder.sh: add REQUIRE_AGENT_TRACE env + --no-require-agent-trace flag,
gate the AGENT_ROWS integrity check behind it.
- m2-ab-run.sh: bump SMOKE_DURATION to 420s; auto-pass --no-require-agent-trace
when AGENT=claude-code; simplify smoke_check to gate only on AO_ROWS>0
(B1 lives in AO-side scm-github, measured by AO trace).
Follow-up tracked as task #39: instrument Claude's hook/tool path or document
that AO_AGENT_GH_TRACE does not cover Claude Code.
* feat(tracker-github): cache issue reads in-process (5 min TTL)
The lifecycle worker polls getIssue/isCompleted repeatedly for the same
issue across a session. Trace data from a 5-session tier-5 bench run
showed the same (repo, issue) pair fetched 64+ times with >97% duplicate
rate — ~744 of 4,059 AO gh calls in 10 minutes were redundant issue views.
Adds an in-process Map<string, CachedIssue> per createGitHubTracker()
instance, keyed by `${repo}#${id}`, TTL 5 min, bounded to 500 entries
(LRU evict-oldest on overflow).
- getIssue: read-through cache, populate on miss
- isCompleted: routes through getIssue (was a separate narrow gh call)
- updateIssue: invalidate the entry before mutating
- createIssue: unchanged, naturally populates via the existing getIssue
- Failures are not cached
Cache lives inside createGitHubTracker so each create() returns an
isolated cache (test isolation comes for free).
Expected reduction: ~744 → ~15 gh issue view calls per tier-5 run.
Tests: 41 existing + 10 new cache tests, all passing.
* feat(scm-github): cache 5 gh pr view callsites with per-method TTLs
The lifecycle worker repeatedly polls each PR for state, summary, reviews,
and review decision. Trace data showed gh pr view was the single largest
AO-side endpoint at 1,280 calls per 5-session tier-5 run with >97% duplicate
rate (e.g. PR #184 polled 86× for --json state alone in 11.5 minutes).
Adds an in-process per-instance cache inside createGitHubSCM(), keyed by
${owner}/${repo}#${prKey}:${method} so different field-sets stay isolated.
Per-method TTLs balance reduction against staleness on decision-influencing
fields:
- resolvePR: 60s (identity metadata only)
- getPRState: 5s
- getPRSummary: 5s (includes state)
- getReviews: 5s
- getReviewDecision: 5s
assignPRToCurrentUser, mergePR, and closePR each invalidate the entire PR
cache for that PR after the mutation, so AO never sees stale state from its
own writes. Failures are not cached.
getCIChecksFromStatusRollup and getMergeability are intentionally NOT cached
here — those need ETag-based revalidation, not blind TTL, and will land
separately.
Expected reduction: ~1,165 of ~1,280 gh pr view calls per tier-5 run.
Tests: 73 existing + 12 new cache tests, all 153 passing.
* feat(scm-github): cache CI checks, mergeability, pending comments, detectPR
Completes the AO-side hot-read caching alongside the prior PR view cache.
All use 5s TTL per the approved policy for decision-influencing fields —
well under one lifecycle poll cycle so state transitions are still seen
next pass.
- getCIChecks (gh pr checks): 5s TTL
- getMergeability (composite pr view + CI + state): 5s TTL on the composite
- getPendingComments (gh api graphql review threads): 5s TTL —
ETag doesn't help on GraphQL per Experiment 2
- detectPR (gh pr list --head BRANCH): 5s TTL, POSITIVE-ONLY.
Empty results are never cached so a freshly created PR is discovered
on the very next poll. The branch-keyed cache entry is invalidated
by mergePR/closePR alongside the number-keyed entries.
Combined with the prior PR view cache, covers the top 6 AO-side gh
operation categories that accounted for ~85% of calls in tier-5 traces.
Tests: 85 existing + 9 new cache tests, all 162 passing.
* experiments(m2): parse REPO from yaml before using it in banner
m2-ab-run.sh referenced $REPO in the header banner before parsing it,
causing 'unbound variable' abort under 'set -u'. Parse it right after
CONFIG_FILE is set.
* test(core): mock full Issue shape in plugin-integration cleanup tests
After tracker-github routed isCompleted() through getIssue() to share
the issue cache, these mocks needed the full Issue shape (number, title,
body, url, state, stateReason, labels, assignees) instead of the narrow
{state} shape that worked when isCompleted made its own --json state call.
* perf(scm-github): tune cache TTLs based on trace replay
Replayed feat run1 + main run2 tier-5 traces (4059 + 1748 rows, 38 min, 5
sessions each) against the shipped cache logic. Three TTLs were materially
under-tuned for the actual lifecycle poll cadence:
- detectPR: 5s → 30s (was 0.5% hit rate; per-branch poll cadence
is ~90s, so 5s caught nothing. 30s catches
intra-cycle bursts when multiple sessions
share a branch. Positive-only stays.)
- getReviewDecision: 5s → 10s (within "10-30s TTL or ETag" policy)
- getPendingComments: 5s → 10s (same policy class)
All three are still well under one poll cycle; freshness contract unchanged
in practice. Other TTLs (5s on state/CI/mergeability, 60s on resolvePR,
5min on issue) hit the targets they were set for and stay as-is.
Replay results before/after:
- feat run1: 53.7% → 57.8% reduction (2179 → 2345 hits of 4059 calls)
- main run2: 47.4% → 52.6% reduction
- Net: ~55% AO-side gh calls eliminated across both traces
Adds experiments/cache-replay.mjs — a counterfactual replay tool that
walks an execGhObserved JSONL trace and simulates per-method cache hits
with the shipped TTLs. Useful as a regression check when tweaking cache
policy.
Tests: 162/162 passing.
* docs(experiments): add cache freshness check runbook
Seven-step manual runbook to validate the cache TTL contract doesn't
cause workflow lag. Covers each cached method with:
- exact gh CLI trigger command
- what to observe in the dashboard / lifecycle log
- pass/fail threshold (TTL + 30s poll cycle)
Companion to experiments/cache-replay.mjs — replay measures how much
we saved, runbook measures whether we lost anything in the process.
* docs(experiments): add Step 2 — consolidate review comment fetching
Single GraphQL call replaces GraphQL + REST for review comments.
Include comment data in agent reaction message to eliminate
agent-side gh read calls. Update future steps.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add duplicate API traffic analysis
Three independent sources hit GitHub API for the same PRs:
1. Dashboard serialize.ts — individual REST calls, no batch, no cache
2. CLI lifecycle manager — batch + guards
3. Web lifecycle manager — same batch + guards, 3s offset
~50% of all API traffic is pure duplication. Dashboard and dual
lifecycle managers are the root causes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add full cache architecture to duplicate traffic analysis
Three independent cache layers across two processes with zero shared
state. Web process creates its own plugin registry, SCM plugin, lifecycle
manager, and dashboard cache — all hitting GitHub independently.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): add shared PR enrichment plan
Persist batch enrichment + review comments to session metadata files.
Dashboard reads from disk instead of making its own GitHub API calls.
Remove web's duplicate lifecycle manager.
Eliminates ~268 calls / 15 min (58% of all traffic). Dashboard data
gets fresher (30s vs 5min). Single writer (CLI lifecycle), web only reads.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): update Step 1 — remove all three fallback paths
Remove fallback in determineStatus(), maybeDispatchCIFailureDetails(),
and maybeDispatchMergeConflicts(). All three follow the same pattern:
batch cache hit → use it, cache miss → skip (wait 30s for next batch).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs(experiments): promote Step 3 (remove dead reviews field) + detail Step 5 (issue caching)
Step 3: Remove reviews(last: 5) from batch query — fetched but never
consumed, reduces GraphQL complexity on every batch call.
Step 5: Persist issue data to session metadata at spawn — eliminates
27 gh issue view calls per 15 min (both processes re-fetch independently).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* feat(core): remove individual REST fallback from lifecycle polling
Remove fallback paths in determineStatus(), maybeDispatchCIFailureDetails(),
and maybeDispatchMergeConflicts() that made individual REST calls when the
batch enrichment cache missed. The batch runs every 30s — a cache miss
means the data arrives on the next cycle, not that it's lost.
Also add populatePREnrichmentCache() call to check() so single-session
checks also use the batch path.
Eliminates ~110 individual pr view/pr checks calls per 15-min window
(24% of all AO-side traffic).
* feat(core): consolidate review comment fetching into single GraphQL call
Add getReviewThreads() to SCM interface — returns all review threads
(human + bot) with isBot flag from a single GraphQL query. Lifecycle
manager splits locally for separate reaction pipelines.
- Eliminates the REST getAutomatedComments() call (40 calls / 15 min)
- Reaction messages now include inline comment data (file, line, author,
body, URL) so agents don't need to re-fetch via gh api
- Default config messages updated to not tell agents to call gh
- getAutomatedComments kept as optional for backward compatibility
* perf(scm-github): remove unused reviews(last: 5) from batch query
The batch query fetched reviews with author, state, submittedAt but
the data was never consumed — only used in a validation check.
The reviewDecision scalar field provides everything AO needs.
Reduces GraphQL complexity cost on every batch call.
* docs(experiments): add post-optimization trace report (Steps 1-3)
5-session, 17-minute trace after removing REST fallback, consolidating
review comments, and removing dead reviews field.
Results: GraphQL 35%/hr (was 41%), REST <1% (was 3%), automated
comment REST calls eliminated. 54% of remaining traffic is redundant
(duplicate lifecycle manager + dashboard individual calls).
* docs(experiments): add trace file gist link to post-optimization report
* feat(core,web): shared PR enrichment — dashboard reads from metadata
CLI lifecycle manager now persists batch enrichment data and review
comments to session metadata files (prEnrichment + prReviewComments
keys). The web dashboard reads from metadata instead of calling
GitHub API.
Changes:
- lifecycle-manager: add persistPREnrichmentToMetadata() after poll,
write prReviewComments in maybeDispatchReviewBacklog()
- serialize: replace enrichSessionPR (6 API calls) with metadata read
- services: stop web lifecycle polling (keep for webhook checks)
- cache: remove prCache (no longer needed)
- routes: remove timeout wrappers and cacheOnly pattern
Eliminates ~237 calls / 15 min (54% of all AO-side traffic).
Dashboard data freshness improves from 5min to 30s.
* fix(web): remove unused beforeEach import in serialize test
* docs(experiments): add final trace report — 56% GraphQL reduction achieved
5-session, 24-min trace after all optimizations including shared
enrichment. GraphQL 905/hr (was 2,072), REST 5/hr (was 168).
Single lifecycle manager confirmed. Dashboard API calls eliminated.
Max sessions before budget exhaustion: ~27 (was ~12).
* docs(experiments): add REST budget breakdown to final report
* fix(core): use storageKey for getSessionsDir in persistPREnrichmentToMetadata
* fix(test): use OpenCodeSessionManager type in plugin-integration tests
* fix(test): update bugbot-comments and auto-cleanup tests for new review API
* fix(web): fix syntax error and missing import from rebase
* docs(experiments): add complete rate-limiting change log and update final report numbers
* fix(web): fix tmux session resolution for legacy wrapped storageKeys
* fix(web): pass tmuxName directly to terminal server instead of reverse-resolving
* perf(core): gate detectPR behind Guard 1 ETag — skip when PR list unchanged
* perf(core): always run Guard 1 for all repos, dedup issue views, include threadId in review messages
* feat(core): add Guard 3 (review ETag), enrich review data with summaries, dedup issue views, gate detectPR for all repos
* fix(web): reuse cached tmuxSessionId on re-open, add 15-session trace report and comparison docs
* perf(scm-github): reduce contexts to first:10, add -i to review GraphQL for rate limit tracing
* feat(core): merge CI details into transition, enrich merge conflict message, reduce batch contexts, add graphqlCost tracing
* chore(experiments): remove working artifacts, keep final reports and reference docs
* chore: remove experiments directory
* refactor: remove getAutomatedComments from SCM interface and all implementations
* fix: address all PR review comments
- gh-trace: make binary resolution async via fs.access (no event loop
blocking, no shell injection), cache mkdir for trace writes, async
fire-and-forget appendFile, document 10MB maxBuffer rationale
- lifecycle-manager: log detectPR failures via observer instead of
silent catch, add getPRState fallback for terminal states
(merged/closed) when batch enrichment cache misses
- scm-gitlab: implement getReviewThreads with bot+human threads and
isBot flag, fixing silent feature regression after
getAutomatedComments removal
- scm-github: clear reviewThreadsCache in invalidatePRCache, document
first:11 CI checks cost budget
- runtime-tmux: use printf+JSON.stringify for PATH export to prevent
shell injection from single quotes
- agent-workspace-hooks: add cache timestamp sanity check, include
--repo in cache keys to prevent cross-repo collisions
- services: document dashboard dependency on CLI polling
- tests: update gh binary path assertions for resolved paths
* fix: address all 17 PR review comments
- gh-trace: use path.delimiter, cache-only-on-success, last HTTP status
line, await writes with warn-once, redact secrets, gate JSON.parse
- agent-workspace-hooks: validate cache keys, redact wrapper trace args,
sha256 cache keys to prevent collisions, 120s TTL ceiling
- session-manager: skip PATH wrappers for claude-code (native hooks)
- types: add deprecation JSDoc for getReviewThreads
- graphql-batch: clear Guard 3 in clearETagCache, re-read ETag on 304,
switch Guard 2 to check-runs endpoint, drop per_page=1 from Guard 3
- Add gh-trace unit tests for extractOperation, redactArgs, parseHttp
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: adil <adil.business4064@gmail.com>
Co-authored-by: iamasx <adilshaikh4064@gmail.com>
Represent missing activity probes as first-class signal states so lifecycle inference only treats valid idle evidence as proof. This prevents false stuck transitions, keeps API/UI lifecycle truth aligned, and makes root monorepo verification deterministic by serializing recursive build and typecheck.
* chore: release 0.2.5
Realign main with npm registry after off-branch publish of 0.2.3/0.2.4.
Bump all 21 linked packages to 0.2.5 and cherry-pick the startup-grace-period
fix for #989 (was in 5e4244a8 but never merged to main).
Also sync non-linked plugin versions (notifier-discord, notifier-openclaw,
scm-gitlab, tracker-gitlab) to their current npm versions.
* Revert "chore: release 0.2.5"
This reverts commit eb17f32834.
* chore: bump all package versions to 0.2.5, remove release workflow
- Bump all 25 packages to 0.2.5 to realign with npm registry
- Update package-version test to expect 0.2.5
- Remove stale .changeset/linear-spawn-branch-name.md
- Delete .github/workflows/release.yml (changesets-based NPM publish)
---------
Co-authored-by: Prateek <karnalprateek@gmail.com>
Co-authored-by: AO Bot <ao-bot@composio.dev>
Renames all npm package scopes from @composio/* to @aoagents/* and
updates GitHub repo references from ComposioHQ/agent-orchestrator
to aoagents/ao throughout the codebase.
- All package.json names and dependencies
- README badges, links, and install instructions
- Documentation references
- Changeset config
- Source code imports and test files
* feat: standardize agent plugins with shared hooks, activity JSONL, and CLAUDE.md
- Add CLAUDE.md with full project context and agent plugin implementation standards
- Extract shared PATH-wrapper metadata hooks into @composio/ao-core (agent-workspace-hooks.ts)
- Backfill Aider + OpenCode: setupWorkspaceHooks, postLaunchSetup, getSessionInfo, getRestoreCommand
- Add recordActivity method to Agent interface for terminal-derived JSONL writing
- Create activity-log.ts in core: appendActivityEntry, readLastActivityEntry
- Lifecycle manager calls recordActivity before getActivityState for agents that implement it
- Upgrade detectActivity in Aider/OpenCode with real terminal prompt/permission patterns
- Upgrade Codex getActivityState to parse JSONL entry types (6 states, up from 2)
- Replace duplicated normalizePermissionMode with shared normalizeAgentPermissionMode from core
- Remove ~200 lines of duplicated shell wrapper code from Codex plugin
- Add git wrapper detection for existing branch switches (parity with Claude Code hooks)
- 484 tests passing across all 4 agent plugins
* chore: unignore CLAUDE.md and AGENTS.md, slim down AGENTS.md to pointer
- Remove CLAUDE.md and AGENTS.md from .gitignore (both should be tracked)
- Slim AGENTS.md from 96 to 30 lines — commands, TL;DR, key files
- Full context now lives in CLAUDE.md; AGENTS.md points there
* fix: eliminate redundant double file read in readLastActivityEntry
Remove the readLastJsonlEntry call that was only used as a null-check,
then immediately discarded before re-reading the same file from scratch.
Now performs a single open + stat + tail-read per poll cycle.
* fix: remove duplicate case-insensitive regex in OpenCode detectActivity
The \(y\)es.*\(n\)o pattern with /i flag was identical to the preceding
\(Y\)es.*\(N\)o/i check — remove the redundant line.
* fix: use zero-initialized buffer and slice to bytesRead in readLastActivityEntry
Replace Buffer.allocUnsafe with Buffer.alloc and slice the result to
actual bytesRead, preventing uninitialized heap data from being parsed
if the file shrinks between stat() and read().
* fix: add staleness cap for waiting_input/blocked and deduplicate recordActivity
- Add ACTIVITY_INPUT_STALENESS_MS (5 min) cap so stale waiting_input/blocked
entries don't keep sessions stuck in needs_input on the dashboard forever.
- Extract checkActivityLogState() into core — shared by aider, opencode, codex.
- Extract classifyTerminalActivity() into core — deduplicates the identical
recordActivity logic across all three plugins.
* fix: prioritize native Codex JSONL over AO activity log in getActivityState
Reorder detection so Codex's native 6-state JSONL (approval_request,
error, tool_call, etc.) is checked first. AO activity JSONL from
terminal parsing is now a fallback only for waiting_input/blocked states
that the native JSONL may not have captured. Previously the AO log was
always fresh (written every poll cycle by recordActivity) and shadowed
the richer native detection entirely.
* fix: restrict AO activity JSONL to waiting_input/blocked only
checkActivityLogState now only returns results for waiting_input/blocked
states. Non-critical states (active/ready/idle) return null, forcing
callers to fall through to their native signals (git commits, chat
history, OpenCode API, Codex native JSONL). This prevents the lifecycle
manager's recordActivity writes (which refresh mtime every poll cycle)
from shadowing richer detection methods and breaking stuck-detection.
* fix: prevent stale idle timestamp in aider and skip flags in git wrapper
- Remove Aider's fallback that returned idle with activityResult.modifiedAt
(always fresh due to recordActivity writes). Now returns null when no
git commits or chat history are found, letting the lifecycle manager
handle stuck-detection correctly.
- Fix git wrapper catch-all case to skip flag arguments (e.g. -B) and
look at $3 for the actual branch name.
* fix: add mtime fallback for empty Codex JSONL and remove unused exports
- When the native Codex session file exists but readLastJsonlEntry returns
null (empty/unparseable), fall back to stat-based mtime detection instead
of losing activity detection entirely.
- Remove unused exports getActivityLogPath and ACTIVITY_INPUT_STALENESS_MS
from @composio/ao-core barrel — they are only used internally.
* fix: opencode activity state detection and CLAUDE.md agent plugin standards
- Fix session ID capture to handle both session_id (snake_case) and
sessionID (camelCase) from OpenCode 1.3.x JSON responses
- Replace broken --command true with -- noop for session creation
(true is not a valid OpenCode command since 1.3.x)
- Add JSONL mtime fallback in getActivityState so active/ready/idle
states work even when findOpenCodeSession returns null
- Rewrite CLAUDE.md activity detection section with the full
getActivityState contract, mandatory JSONL mtime fallback pattern,
and 8 required tests every agent plugin must implement
* fix: opencode --command true flag and activity JSONL mtime staleness
- Fix opencode getLaunchCommand to use `--command true` instead of `-- noop`
(aligns with test expectations and opencode CLI docs)
- Fix checkActivityLogState to use entry.ts instead of file mtime for
staleness checking — recordActivity refreshes mtime every poll cycle,
which prevented stale waiting_input/blocked entries from being detected
- Fix opencode getActivityState fallback to use entry state directly
instead of re-deriving from mtime, which always returned "active"
because recordActivity refreshes the file every cycle
- Update tests to reflect new entry-state-based fallback behavior
* fix: deduplicate recordActivity writes and restore mtime-based fallback
recordActivity was writing to the JSONL every poll cycle (~30s), which
kept refreshing the file mtime and prevented the JSONL mtime fallback
in getActivityState from ever reaching "ready" or "idle".
Fix: skip writes when the state hasn't changed and the last entry is
<20s old. This keeps mtime fresh during active work (writes every
20-30s, within the 30s activeWindow) but lets it age naturally when
the agent goes quiet.
Also restores the mtime-based age classification in the JSONL fallback
(active/ready/idle by mtime age) instead of returning the entry state
directly, which was always "active" since that's what recordActivity
writes.
Applied to both OpenCode and Aider plugins. Updated CLAUDE.md with
the dedup pattern and rationale.
* fix: align integration tests with opencode `-- noop` launch command
Update 9 test expectations from `--command true` to `-- noop` to match
the reverted getLaunchCommand implementation.
* fix: add JSONL mtime fallback to Aider getActivityState
When git commits and chat history are both unavailable (e.g. early
session startup), Aider's getActivityState now falls back to the AO
activity JSONL mtime for active/ready/idle classification — matching
OpenCode's existing step 3 fallback. Previously it returned null,
leaving the dashboard with no activity signal.
* fix: add write deduplication to Codex recordActivity
Add the same dedup logic that Aider and OpenCode already have — skip
writes when the state hasn't changed and the last entry is recent
(<20s). Prevents unbounded file growth and stale mtime refreshes.
* refactor: extract shared recordTerminalActivity into core
Move the duplicated recordActivity logic (classify + dedup + append)
from all three plugins into a shared `recordTerminalActivity` function
in `@composio/ao-core/activity-log`. Each plugin's `recordActivity`
is now a thin wrapper that delegates to the shared function.
Add core tests for classifyTerminalActivity, checkActivityLogState,
and recordTerminalActivity (10 tests).
* fix: validate JSONL entry fields, handle invalid dates, consistent operators
- Validate required fields (ts, state, source) before casting parsed
JSON to ActivityLogEntry — prevents malformed entries from propagating
- Guard against invalid Date parsing in checkActivityLogState — returns
null instead of comparing against NaN
- Use <= instead of < in Aider chat-history threshold comparisons to
match OpenCode and Aider's own JSONL fallback path
* fix: use --command true for opencode run, validate sed key parameter
- Replace `-- noop` with `--command true` in opencode getLaunchCommand
so the bootstrap uses a valid command
- Validate metadata key against [a-zA-Z0-9_-]+ in the git wrapper's
update_ao_metadata to prevent sed metacharacter injection
* fix: extract DEFAULT_ACTIVE_WINDOW_MS constant, clarify Codex recordActivity
- Extract magic number 30_000 into DEFAULT_ACTIVE_WINDOW_MS constant in
core types, used by Aider and OpenCode for active/ready thresholds
- Clarify in CLAUDE.md that Codex implements recordActivity as a safety
net for when its native JSONL is missing/unparseable, not redundantly
* fix: add JSONL mtime fallback to Codex getActivityState
When native Codex session file is missing but AO JSONL has data,
derive active/ready/idle from JSONL mtime instead of returning null.
Matches the fallback pattern already in Aider (step 4) and OpenCode
(step 3).
* fix: validate ActivityState and source values when parsing JSONL entries
Validate that `state` is one of the known ActivityState values and
`source` is "terminal" or "native" before constructing the entry.
Construct the entry explicitly instead of using unsafe double cast.
* fix: wrap getOutput + recordActivity in try-catch to protect getActivityState
If runtime.getOutput() throws (e.g. tmux unresponsive), the error
previously propagated to the outer catch, skipping getActivityState
entirely. Now the entire recordActivity preamble is wrapped in its
own try-catch so getActivityState always runs.
* chore: add .ao/ to gitignore
* chore: clean up gitignore comment
* test: add coverage for activity-log and agent-workspace-hooks
- activity-log: test readLastActivityEntry (missing file), invalid
entry.ts in checkActivityLogState, blocked state path
- agent-workspace-hooks: test buildAgentPath (dedup, defaults,
ordering), setupPathWrapperWorkspace (create/skip wrappers,
AGENTS.md create/skip)
Raises diff coverage from 39% toward 80% threshold.
* test: add real file I/O tests for readLastActivityEntry and recordTerminalActivity
Test readLastActivityEntry with actual temp files: valid entries,
empty file, invalid JSON, invalid state, missing fields, multi-line.
Test recordTerminalActivity dedup logic and actionable state bypass.
* fix: add activeWindow threshold to Codex native JSONL state detection
Action entries (tool_call, user_input, exec_command) now use the 30s
active window: <=30s is "active", 30s-5min is "ready", >5min is "idle".
Previously these skipped "ready" entirely, going straight from "active"
to "idle" at the 5min threshold.
* fix: add activeWindow threshold to Claude Code native JSONL state detection
Same fix as Codex — action entries (user, tool_use, progress) now use
the 30s active window for consistent 3-state classification across
all agent plugins: <=30s active, 30s-5min ready, >5min idle.
* fix: handle truncated JSONL, add exited state, fix session lookup and dedup window
- readLastActivityEntry: increase tail buffer to 4KB, skip truncated
first line when reading from offset, try lines from end on parse error
- Add "exited" to valid ActivityState set in JSONL validation
- findOpenCodeSession: pick most recently updated session when multiple
title matches exist, preventing stale session binding
- Increase dedup window from 20s to 60s so mtime can age past the 30s
active window between writes, allowing "ready" state to be reached
* fix: use entry state for JSONL fallback, write AGENTS.md to .ao/, revert dedup to 20s
- Replace mtime-based active/ready/idle derivation in all 3 plugin
fallbacks with direct entry.state + entry.ts usage. This eliminates
the fundamental conflict between write deduplication and mtime
freshness — the entry already has the correct detected state.
- Revert dedup window to 20s (purely I/O optimization, no longer
affects state detection since mtime is not used)
- Write AO session context to .ao/AGENTS.md (gitignored) instead of
modifying the repo-tracked AGENTS.md, preventing dirty worktree state
* fix: reorder Codex fallback chain so AO JSONL is checked before stat mtime
When native JSONL exists but can't be parsed, the stat() fallback
previously returned early, skipping AO JSONL waiting_input/blocked
detection and the ready state. Now it falls through to AO JSONL first,
then uses stat mtime as a last resort with proper 3-state classification.
* fix: re-validate canonicalized ao_dir against trusted roots
After resolving symlinks with pwd -P, re-check real_ao_dir against
the trusted root allowlist. Prevents paths like /tmp/../../home/user
from passing the pre-canonicalization check then escaping to arbitrary
directories after symlink resolution.
* fix: update tests for .ao/AGENTS.md location and remove unused vi import
- Update Codex setupWorkspaceHooks tests to expect .ao/AGENTS.md
instead of workspace root AGENTS.md
- Remove unused vi import from activity-log test (lint error)
* fix: add age-based decay to JSONL entry fallback via getActivityFallbackState
Extract getActivityFallbackState in core — reclassifies entry state
based on entry.ts age (active→ready→idle) so old entries don't stay
as "active" forever when recordActivity stops being called. All three
plugins now use this shared helper for their JSONL fallback paths.
* fix: apply staleness cap to actionable states in getActivityFallbackState
Stale waiting_input/blocked entries (older than ACTIVITY_INPUT_STALENESS_MS)
are now treated as idle in the fallback, preventing them from bypassing
the staleness filtering in checkActivityLogState.
* fix: respect entry state as ceiling in getActivityFallbackState
Age-based decay can only demote (active→ready→idle), never promote.
A fresh "idle" entry stays "idle" instead of being reclassified as
"active" — the detected state from terminal output is authoritative.
* docs: update CLAUDE.md to match current activity detection implementation
- Replace inline recordActivity dedup example with recordTerminalActivity delegation
- Replace mtime-based fallback example with getActivityFallbackState
- Update step 4 description: entry state + age-based decay, not mtime
- Add new core exports to utilities section
- Document .ao/AGENTS.md location and setupPathWrapperWorkspace
- Update required test list for entry-based fallback
* fix: skip metadata helper rewrite when version marker matches
Move ao-metadata-helper.sh write inside the needsUpdate check so it's
only rewritten when wrapper scripts are outdated, not on every call.
* fix: update Codex test for metadata helper skip when version matches
Metadata helper is now inside the needsUpdate check, so when the
version marker matches, no wrapper writes occur (including helper).
- Fix autoDetectProject path matching: expand ~ before comparing project
paths to cwd, so `path: ~/my-repo` matches `/Users/user/my-repo`
- Fix addProjectToConfig: detect duplicate directory names and auto-suffix
instead of silently overwriting existing project entries
- Fix agent detect() in all 4 plugins: replace `which` (Unix-only) with
direct `--version` invocation for cross-platform compatibility
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Change != null to !== undefined && !== null in caller-context.ts and
session-manager.ts (3 locations) to satisfy eqeqeq lint rule
- Add displayName field to all 4 agent plugin test manifest assertions
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Replace require() with execFileSync in all 4 agent plugin detect()
functions — fixes ReferenceError in ESM-only environments
- Remove non-existent -p flag from orchestrator prompt spawn/batch-spawn
docs — prevents orchestrator agent from generating broken commands
- Return actual port from runStartup() and pass to register() — fixes
running.json storing wrong port when auto-escalation picks a free port
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Reduces onboarding to `npm install -g @composio/ao && ao start`.
- Absorb init and add-project logic into `ao start` with auto-config
creation, environment detection, and project type detection
- Add single-instance tracking via running.json with already-running
interactive menu (human) and info+exit (agent)
- Add caller context detection (human/orchestrator/agent) via TTY and
AO_CALLER_TYPE env var
- Add plugin-based agent runtime detection — no hardcoded binary paths,
each plugin exports detect() and displayName
- Simplify `ao spawn` to just `ao spawn <issue>` with auto-detected
project (backward compat: `ao spawn <project> <issue>` still works)
- Set AO_CALLER_TYPE, AO_PROJECT_ID, AO_CONFIG_PATH, AO_PORT env vars
on all spawned sessions (worker, orchestrator, restore)
- Add `ao config-help` subcommand for config schema reference
- Deprecate `ao init` to thin wrapper, fully remove `ao add-project`
- Register/unregister in running.json on start/stop
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(lifecycle): implement stuck detection using agent-stuck threshold
The agent-stuck reaction config supported a threshold field (e.g.
"10m"), but determineStatus() never returned "stuck" — there was no
code path that consumed the threshold or transitioned sessions based
on idle time. Sessions would stay parked at pr_open/working forever
even when the agent had been idle for hours.
Added idle-time check in determineStatus(): when getActivityState()
reports "idle" or "blocked" with a timestamp, compare the idle
duration against the agent-stuck.threshold config. If exceeded,
return "stuck" so the reaction system can fire notifications.
Also removed the priority !== "info" guard on transition
notifications, so all priority levels (including info) are routed
through notificationRouting. This lets the config control which
notifiers receive each priority level, rather than silently dropping
info-level transition events.
* fix(lifecycle): add post-PR stuck detection as safety net
The original stuck check in step 2 (before PR checks) can be bypassed
when getActivityState() returns null (session file not found, cache miss,
I/O failure). When this happens, the code falls through to the PR path
which returns 'pr_open' without ever checking idle duration.
Fix: extract isIdleBeyondThreshold() helper and call it in three places:
1. Step 2: before PR checks (fast path, catches most cases)
2. Step 4b: after PR checks return 'pr_open' (safety net)
3. Step 5: after all checks, for agents that finish without a PR
This ensures stuck detection fires even when the JSONL activity detection
fails to return idle state. Sessions can no longer get permanently stuck
at 'pr_open' when the agent has been idle beyond the threshold.
Also removes the debug console.error calls from the previous commit.
* fix(lifecycle): treat reviewDecision 'none' as approved for merge readiness
PRs with no required reviewers never reached 'mergeable' status because
getReviewDecision returned 'none', which was not handled. The lifecycle
poll fell through to 'review_pending' or the default, so merge.ready
never fired and the approved-and-green reaction never triggered.
Also: skip stuck short-circuit when session has an open PR so merge
readiness checks in step 4 can still run. Without this, idle agents
with open PRs get stuck status and never transition to mergeable.
Closes composio#0 (internal fix)
* fix(opencode): classify activity state from session timestamps
* test(lifecycle): cover opencode idle-threshold stuck transition
* fix(lifecycle): preserve global stuck threshold with project overrides
* fix(lifecycle): run PR auto-detection before stuck transitions
---------
Co-authored-by: Harsh <harshb012@gmail.com>
* fix: capture exact OpenCode session ID from JSON stream to prevent orphan sessions
- Primary: Extract session_id from opencode run --format json step_start event
- Fallback: Title-based match with newest-first sorting for delayed visibility
- Core: Add atomic reserveSessionId check in spawnOrchestrator to prevent race conditions
- Prevents orphan sessions when multiple spawns use same title or delayed discovery
(cherry picked from commit 42cc0cfa1a)
* fix: capture OpenCode session id from stream output
(cherry picked from commit 42b3bf3ba6)
* fix: respawn orchestrators when stale metadata is left behind
(cherry picked from commit e519628017)
* fix: align respawn guard and opencode test expectation
* fix: remove unused helper from opencode launch path
* fix: remove unused function and align tests with session capture format
- Remove unused buildSessionLookupScript function from opencode agent
- Update integration tests to expect --format json flag in launch commands
- Fix assertions for exec opencode --session wrapper format
* fix: address Bugbot findings on PR #366
- Fix orphaned runtime leak when reuse strategy finds alive runtime but
get() returns null (now destroys the orphaned runtime)
- Restore session ID format validation in fallback script with
isValidId regex check
- Add regression tests for both fixes
* fix: address additional Bugbot findings on PR #366
- Add timestamp helper to fallback sort to avoid NaN from invalid dates
- Add session ID type/format validation to primary capture script
- Add tests for both robustness improvements
* fix: address bugbot follow-ups in send and opencode discovery
* fix(test): update stale integration test expectation for opencode bootstrap
The getLaunchCommand implementation now uses a robust bootstrap pattern
that captures the session ID between 'opencode run' and 'exec opencode --session'.
Updated test to check both parts separately instead of expecting a contiguous
string that no longer exists.
* fix: avoid NaN in sort comparator when both timestamps are missing
The comparator (b.updatedAt ?? -Infinity) - (a.updatedAt ?? -Infinity)
produces NaN when both timestamps are missing because -Infinity - (-Infinity)
is NaN in IEEE 754. A comparator returning NaN violates ECMA-262's
'consistent comparison function' requirement, making sort results
implementation-defined.
Fixed by adding equality check before subtraction:
- If both timestamps are equal (including both -Infinity), return 0
- Otherwise return the difference
* docs: add design research artifacts — briefs, token reference, screenshots
Comprehensive design research package for the ao dashboard, session
detail page, and orchestrator terminal. Produced via competitive analysis
of 14 products (Linear, Vercel, Railway, Fly.io, Inngest, WandB, LangSmith,
Supabase, and more) + Playwright CSS extraction from live sites + full
codebase audit.
Artifacts:
- docs/design/design-brief.md Main design brief (v2, Playwright-updated)
- docs/design/session-detail-design-brief.md /sessions/[id] design spec
- docs/design/orchestrator-terminal-design-brief.md Orchestrator page spec
- docs/design/token-reference.css Drop-in CSS replacement for globals.css
- docs/design/competitive-analysis-raw.md Raw research notes, all 14 sites
- docs/design/design-brief-v1.md Original text-only brief (pre-Playwright)
- docs/design/README.md Index + research methods summary
- docs/design/screenshots/linear-homepage.png Playwright-captured screenshot
- docs/design/screenshots/railway-homepage.png Playwright-captured screenshot
Key findings:
- Linear CSS token values verified via Playwright (body bg #08090A, accent
#7070FF, Berkeley Mono monospace, type scale, radius, transitions)
- Recommended palette: #0C0C11 base (blue-cast dark vs current GitHub #0d1117)
- Highest-impact change: load Inter Variable via next/font/google
- Orchestrator terminal needs visual differentiation (violet accent, status strip)
- token-reference.css is ready to drop into packages/web/src/app/globals.css
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(web): redesign dashboard, session detail, and orchestrator terminal
Implements a cohesive dense dark-mode design system across all three main views.
- New color token palette: #0c0c11 base, #141419 surface, #1c1c25 elevated
- Accent blue #5b7ef8, status semantics (ready/error/attention/working/idle/done)
- Violet accent #a371f7 reserved for orchestrator
- Inter Variable + JetBrains Mono loaded via next/font with CSS variables
- activity-pulse keyframe for live agent dots
- AttentionZone header: dot + label + flex divider + count pill + chevron
- Sessions laid out in responsive 1→2→3 column grid
- Solid green merge button (translateY hover), no confirm() dialog
- Breadcrumb nav: ← Agent Orchestrator / {session-id} [orchestrator badge]
- CSS 8×8px activity dot with pulse animation replaces emoji labels
- Merge-ready state: green-bordered banner with checkmark icon
- Orchestrator sessions show zone counts strip (merge/respond/review counts)
- xterm.js dark theme (#0a0a0f bg, #d4d4d8 fg, full 16-color ANSI palette)
- variant prop: "agent" (blue cursor) vs "orchestrator" (violet cursor)
- Dynamic height prop instead of fixed 600px; fullscreen toggle with SVG icons
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* refactor(web): strip rainbow stats, clean header, IBM Plex Sans typography
- Replace Inter with IBM Plex Sans (technical tool aesthetic, distinctive numerics)
- Replace 4-color big-number stats bar with a single compact inline status line
in the header: "35 sessions · 1 working · 9 PRs" — no decorative colors
- Remove the two-tone "Agent (blue) Orchestrator (white)" title — just "Orchestrator"
- Remove ClientTimestamp (useless) — replaced by orchestrator nav link
- Zone headers: colored dot only (semantic), neutral uppercase label, plain count
— removes the rainbow-colored label text that read as a widget template
- Add subtle radial gradient glow at top of page for depth
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(web): kanban layout, amber accent, full-width, bigger stats
- Switch accent from blue (#5b7ef8) to amber/gold (#d18616) throughout
- Replace grid layout with horizontal Kanban columns for active zones
(merge, respond, review, pending, working), Done stays full-width below
- Remove max-w-[1100px] constraint — full viewport width
- Header stats numbers 20px bold (was 12px), orchestrator link is now a
visible bordered button
- AttentionZone gains variant="column" for Kanban mode (compact header
with count pill, vertical card stack)
- Update all hardcoded rgba(91,126,248,...) in SessionCard to amber
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): layout, alert sizing, column order, button feel
- Kanban column order: working→pending→review→respond→merge
(left = in progress, right = ready to ship)
- Columns use flex-1 min-w-[200px] to fill available width
instead of fixed 260px leaving half the page empty
- Alert badges: inline-flex wrapper prevents stretching to full
row width when wrapping
- Terminal button: add bg-subtle fill so it reads as a button
- PR number (#91): remove opaque pill background, now plain
amber text link — clearly a hyperlink
- Merge PR button: pt-0.5 spacer above the action area
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): don't cache rate-limited partial PR data
When GitHub rate limits fire, enrichSessionPR was caching the
bad partial data (0 additions, CI failing) for 60 seconds, causing
the dashboard to show incorrect data for the full TTL window.
- Skip cache write when majority of API calls failed
- Downgrade console.error → console.warn (this is handled/expected)
The next page refresh will retry live API calls, so data recovers
as soon as the rate limit window resets.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(web): graceful GitHub API rate limit handling in UI
When the GitHub plugin hits rate limits, the dashboard now:
- Shows a single amber banner: "GitHub API rate limited — PR data
(CI status, review state, sizes) may be stale. Will retry on
next refresh."
- Hides CI badge, review decision, and size pill on PR cards
(they'd show wrong values: +0 -0 XS, CI failing)
- Shows a subtle "⚠ PR data rate limited" note on affected cards
instead of misleading alert badges
- Skips CI/review/conflict-based attention zone classification
for rate-limited PRs (prevents sessions moving to Review due
to phantom "CI failing" from the fallback value)
- Doesn't cache partial rate-limited data so next refresh retries
live API calls as soon as the rate limit window resets
What still works when rate limited:
- Session ID, title, branch, PR number/link
- Session activity status (working/spawning/etc.)
- Merge button if mergeability was already cached
- Restore/terminate/send actions
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(web): dashboard redesign — glassmorphism, Kanban, rate-limit handling, perf fix
Design:
- Kanban layout: active zones as flex columns (working→pending→review→respond→merge),
Done as full-width grid below
- GitHub dark color palette (main's tokens) with glassmorphic card surfaces
(rgba bg + backdrop-blur) and subtle blue/violet body gradient
- Activity state shown as labeled pill (● active / ● idle etc.) instead of bare dot
- Session card: title on its own row, larger font, inline-flex alert badges (no stretch)
- PR number rendered as plain accent link, not a blue pill badge
- Terminal button has background fill to feel like a button
- Info circle icon replaces alarming warning triangle for rate-limit indicators
- "1 working" → "1 active" in header stats
- PR table constrained to max-w-[900px] and centered
- Orchestrator session no longer uses purple accent
Rate limiting:
- isPRRateLimited() helper; getAttentionLevel() skips PR classification when limited
- Rate-limited banner in Dashboard; suppressed CI/size/review badges in PRStatus
- SessionCard shows subtle "PR data rate limited" indicator; getAlerts() returns []
- serialize.ts: rate-limited enrichment results cached for 5 min (not 60s) to stop
retrying 168 failing API calls every minute
Performance:
- page.tsx: 4s hard timeout on PR enrichment — serves stale data fast instead of
blocking SSR for 75s under rate limiting
- cache.ts: TTLCache.set() accepts optional ttl override for per-entry control
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: suppress stale size/CI/review in PR table when rate limited
PRTableRow now shows "—" for size, CI, and review columns when GitHub
API is rate limited, matching the card view which already hides these.
Prevents misleading "+0 -0 XS" size and "needs review" labels from the
default fallback values.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: 3D card effect with depth shadow and top-edge shine
Cards now clearly pop against the dark background:
- Solid gradient bg (rgba(28,36,47) → rgba(18,23,31)) instead of
near-invisible rgba(22,27,34,0.8) surface
- Layered box-shadow: contact shadow + diffuse depth + inset top highlight
that simulates light hitting the card's top edge (the "shine")
- Hover: card lifts 2px with deeper shadow
- Merge-ready: green-tinted bg with green ambient glow + stronger lift on hover
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: restore text legibility inside session cards
The darker solid card gradient made muted/secondary text nearly
invisible — #484f58 (text-muted) had only ~2:1 contrast on the
new card bg. Override the color tokens locally within .session-card
to GitHub's established dark-mode legibility values:
--color-text-muted: #484f58 → #656d76 (3.8:1 on card bg)
--color-text-secondary: #7d8590 → #8b949e (6.2:1 on card bg)
--color-text-tertiary: #484f58 → #656d76
Scoped to .session-card so the rest of the UI is unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: address bugbot comments — fonts, review zone, ActivityDot, orchestrator btn
- layout.tsx: add IBM Plex Sans weight 700 (was missing, font-bold falling
back to 600)
- DirectTerminal.tsx: use "IBM Plex Mono" instead of unloaded "JetBrains Mono"
- SessionDetail.tsx: add review zone to OrchestratorStatusStrip (was omitted,
sessions with CI failures were invisible in the strip)
- ActivityDot.tsx: extract shared component, remove duplicate implementations
in SessionCard.tsx and SessionDetail.tsx
- Dashboard.tsx: redesign orchestrator button with 3D glass style matching
card aesthetic (blue-tinted bg, depth shadow, hover lift)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): lint — eqeqeq, duplicate import, unused var
- ActivityDot.tsx: != → !== (eqeqeq rule)
- PRStatus.tsx: merge duplicate @/lib/types imports into one
- SessionCard.tsx: remove unused activityIcon import
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(web): elevate session detail + orchestrator page design
- Nav: glass backdrop-blur effect with chevron back link
- Header: detail-card 3D treatment with left-border accent keyed to activity color
- Meta chips: bordered pill style with subtle bg instead of flat text
- Status tag: pill badge for status instead of plain text
- PR card: detail-card 3D treatment, border-color reflects PR state
- PR merged badge: purple pill instead of gray text
- Unresolved count: red pill badge in section header
- Blockers section: renamed "Issues" → "Blockers"
- Terminal section: colored bar indicator instead of plain label
- Orchestrator status strip: total agent count + per-zone colored pills
- globals.css: add .nav-glass and .detail-card classes
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): fetchZoneCounts parses body.sessions, delayed 2s to avoid contention
The /api/sessions endpoint returns `{ sessions: [...] }` not a bare array.
fetchZoneCounts was treating the whole response object as an array, so
zone counts were always zero on the orchestrator detail page.
Also delays the initial fetchZoneCounts call by 2s so it doesn't contend
with the session fetch on page load (both hit /api/sessions which is slow
when GitHub enrichment is running).
Also includes: Playwright kill-Chrome-for-Testing tip in CLAUDE.md,
toned-down detail-card shadow in globals.css.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* perf+test(web): cache-first PR enrichment, skip exited sessions, fix component tests
Performance improvements:
- enrichSessionPR() now accepts cacheOnly option and returns boolean
- /api/sessions/[id]: serve from cache immediately, only block on first load
- /api/sessions: skip PR enrichment for EXITED sessions (no longer changing)
- cache: increase default TTL from 60s to 5 minutes
Test fixes (match redesigned SessionCard + AttentionZone):
- "restore session" (header) → "restore"; expanded panel still shows "restore session"
- "merge PR #N" → "Merge PR #N" (capital M)
- "CI status unknown" → "CI unknown"
- "ask to fix CI" / "ask to fix CI" → "ask to fix"
- "terminate session" → "terminate"
- Zone labels: RESPOND/WORKING/DONE → Respond/Working/Done (CSS uppercase is visual only)
- "working" zone no longer collapsed by default; collapse tests now use "done" zone
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat(core): ActivityDetection with timestamp propagation
- Add ActivityDetection interface { state, timestamp? } to types.ts
- Agent getActivityState() returns ActivityDetection | null instead of
ActivityState | null, allowing timestamp from JSONL mtime to propagate
- session-manager updates session.lastActivityAt when detected.timestamp
is more recent — fixes "active 22h ago" showing stale timestamps
- Update all agent plugins (claude-code, aider, codex, opencode) to
return ActivityDetection objects
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): dismissible rate limit banner + 60min rate-limit cache TTL
- Add X dismiss button to GitHub API rate limit banner in Dashboard.tsx
so it can be closed during demos
- Extend rate-limited PR cache TTL from 5min to 60min — GitHub GraphQL
rate limits reset hourly, no point retrying every 5 minutes
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): address Cursor Bugbot review comments on PR #125
- Dashboard StatusLine: active sessions count now uses var(--color-status-working)
(blue) instead of neutral text color, matching the design system semantics
- SessionCard: isReadyToMerge now guards against rate-limited state — a card
with stale cached mergeability data won't show green merge-ready styling
- DirectTerminal: add `variant` to useEffect dependency array (was missing,
causing stale cursor/selection colors if variant changed after mount)
- agent-aider: include `timestamp: chatMtime` in all ActivityDetection returns,
matching the pattern used by agent-claude-code (enables accurate lastActivityAt)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(ci): resolve lint, typecheck, and test failures
Lint:
- Remove unused parseJsonlFile function (superseded by parseJsonlFileTail)
- Remove dead lastLogModified stat() call in getSessionInfo (field was
removed from AgentSessionInfo but the filesystem read was left behind)
Typecheck + Tests (ActivityDetection):
- session-manager.test.ts: update mocks to return { state: "active" } /
{ state: "idle" } instead of bare strings — getActivityState() returns
ActivityDetection | null, not ActivityState | null
- integration tests (aider, claude-code, codex, opencode): update imports
from ActivityState → ActivityDetection, variable types, comparisons
(activityState?.state !== "exited"), and assertions (?.state).toBe()
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: parseJsonlFileTail uses readFile for small files; enrich exited sessions with PRs
- parseJsonlFileTail now calls stat() then readFile() for files smaller than
maxBytes, falling back to open()/handle.read() only for large files. This
fixes the test infrastructure (which mocks readFile but not open) and also
fixes a scope bug where `offset` was declared inside an inner try block but
referenced outside both try blocks.
- Math.max(0, NaN) returns NaN not 0, so size must default to 0 when stat
returns a mock without a size field: `const { size = 0 } = await stat(...)`.
- Update activity-detection.test.ts: getActivityState() now returns
ActivityDetection objects, so tests use (await ...)?.state comparisons.
- Remove stale lastLogModified test (field was removed from AgentSessionInfo).
- Remove EXITED skip guard from api/sessions/route.ts: exited sessions can
still have open, merge-ready PRs that need enrichment on the dashboard.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: comprehensive code review fixes — tests, timestamps, UI correctness
Address gaps identified in code review of the ActivityDetection PR:
Core / Session Manager:
- Add `timestamp` to all `{ state: "exited" }` returns in all 4 agent plugins
(claude-code, aider, codex, opencode) using consistent `exitedAt = new Date()` pattern
- Add 2 new session-manager tests: timestamp propagation when detection timestamp
is newer, and no-downgrade when detection timestamp is older
- Fix `parseJsonlFileTail` lint error: remove useless `= 0` initializer (value was
always overwritten before use; catch block returns early)
Web package — tests:
- Fix 3 `api-routes.test.ts` failures: `sessionsGET()` needs a Request object since
the route reads `request.url` for `?active=true` query param
- Fix `serialize.test.ts` rate-limit test: spy on `console.warn` (what the code uses)
not `console.error`
- Add 5 `ActivityDot` component tests covering all activity states, unknown states,
null activity, and dotOnly mode
Web package — UI correctness:
- Fix `relativeTime()` in SessionDetail to guard against invalid/empty ISO strings
- Fix timer Map leak: add `timersRef.current.clear()` in cleanup effect after forEach
- Add `encodeURIComponent` to sessionId in message fetch URL
Server — race condition fix:
- Guard `activeSessions.delete` in pty.onExit, ws.on("close"), and ws.on("error")
against stale handlers deleting a newly-registered session with the same ID.
Fixes flaky integration test where afterEach's pty.kill() fired asynchronously
after the next test had already set up a new session with the same session ID.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix(web): narrow PREnrichmentData types to eliminate unsafe casts in serialize
PREnrichmentData.ciStatus and .reviewDecision were typed as string,
requiring unsafe `as` casts when reading from cache into DashboardPR.
Narrow them to the same literal union types used by DashboardPR, making
the casts unnecessary. Also narrow ciChecks[].status to match CoreCICheck.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: activity detection — fix path encoding, use tail -1 for JSONL
Two tightly coupled infrastructure fixes:
- Fix toClaudeProjectPath(): leading `/` becomes `-` (not stripped),
matching Claude Code's actual project directory naming convention.
- Replace manual 4KB buffer read in readLastJsonlEntry() with
`tail -1` + JSON.parse — handles any file size, any line length,
and eliminates the truncated-line edge case entirely.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add "ready" state, return null when unknown, remove dead code
Behavioral changes to activity detection:
- Add "ready" to ActivityState — separates "alive at prompt" from
"idle/stale". Configurable via readyThresholdMs (default 5 min).
- Agent plugins return null when they can't determine activity
(no workspace, no JSONL, no per-session tracking). Session manager
preserves existing activity instead of overwriting with a guess.
- Remove isProcessing() from Agent interface — zero callers in
production code, fully superseded by getActivityState().
- Remove extractLastMessageType() from claude-code — the field it
populated (lastMessageType) was only consumed by the old inline
CLI mapping, which is now replaced by plugin delegation.
- CLI status delegates to agent.getActivityState() (single source
of truth) with metadata fallback when plugin returns null.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* test: comprehensive activity detection coverage
- activity-detection.test.ts: 42+ tests covering path encoding,
getActivityState edge cases (exited/null/fallback), real Claude Code
JSONL types, agent interface spec types, staleness thresholds,
JSONL file selection, and realistic session sequences.
- status.test.ts: plugin delegation tests — verifies CLI uses
agent.getActivityState() as single source of truth, passes
readyThresholdMs from config, falls back to metadata on null/throw.
- Integration tests: updated type expectations for null returns from
codex, opencode, and aider; added "ready" to valid state lists.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: flaky Linear integration test + missing ready label in SessionDetail
Linear API has eventual consistency — updateIssue state changes don't
propagate instantly. Poll with retries instead of asserting immediately.
Also adds "ready" entry to SessionDetail activityLabel map (was missing,
causing fallback to dim/unstyled rendering).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: pure Node.js readLastJsonlEntry, use pollUntilEqual for Linear test
Replace `tail -1` with pure Node.js implementation that reads backwards
from end of file in 4KB chunks. No external binary dependency — works
on any platform.
Fix flaky Linear integration test by using the existing pollUntilEqual
helper instead of an inline retry loop. Linear API has eventual
consistency; pollUntilEqual retries for up to 5s with 500ms intervals.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: use tail -1 for readLastJsonlEntry, add real-data integration test
Replace over-engineered pure Node.js backward-reading implementation with
simple `tail -1` via execFile. The codebase already shells out to tmux,
git, and ps everywhere — tail is no different.
Add integration test that validates toClaudeProjectPath() and
readLastJsonlEntry() against real ~/.claude/projects/ data on disk.
No API key needed — just requires Claude to have been run once.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat: implement seamless onboarding with enhanced documentation
- Add comprehensive README.md (18KB) with quick start, core concepts, and FAQ
- Add detailed SETUP.md (16.5KB) with prerequisites, integration guides, and troubleshooting
- Add examples/ directory with 5 ready-to-use config templates:
- simple-github.yaml: Minimal GitHub setup
- linear-team.yaml: Linear integration
- multi-project.yaml: Multiple repos
- auto-merge.yaml: Aggressive automation
- codex-integration.yaml: Using Codex agent
- Add environment detection (git repo, remote, branch, auth status)
- Auto-fill prompts with smart defaults from detected environment
- Add prerequisite validation (git, tmux, gh CLI)
- Show actionable next steps and warnings
- Parse owner/repo from git remote automatically
- Detect LINEAR_API_KEY and SLACK_WEBHOOK_URL in environment
- Prompt for Linear team ID when Linear tracker selected
- Format all files with Prettier for consistency
Reduces onboarding time from 30+ minutes to ~5 minutes:
1. Install CLI: `npm install -g @composio/ao-cli`
2. Run init: `ao init` (auto-detects everything)
3. Spawn agent: `ao spawn my-project ISSUE-123`
Users no longer need to:
- Manually parse git remote URLs
- Look up current branch names
- Remember YAML syntax
- Search for Linear team IDs
- Debug missing prerequisites
- ✅ pnpm build - All packages compile
- ✅ pnpm typecheck - No TypeScript errors
- ✅ pnpm lint - No new linting issues
- ✅ pnpm format - All files formatted
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* docs: update installation instructions to reflect npm not yet published
Package is not published to npm yet, so users must build from source.
Updated README.md and SETUP.md to:
- Make 'build from source' the primary installation method
- Add note that npm publishing is coming soon
- Include pnpm as a prerequisite
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add ao init --auto --smart for zero-config setup
Implements intelligent config generation with project type detection.
## What's New
### ao init --auto
- Zero prompts - auto-generates config with smart defaults
- Detects: git repo, remote, branch, languages, frameworks, tools
- Generates project-specific agentRules based on detected tech stack
### Project Detection
- Languages: TypeScript, JavaScript, Python, Go, Rust
- Frameworks: React, Next.js, Vue, Express, FastAPI, Django, Flask
- Tools: pnpm workspaces, test frameworks
- Package managers: pnpm, yarn, npm
### Rule Templates
Created templates for:
- base.md - Universal best practices
- typescript.md - TS strict mode, ESM, type imports
- javascript.md - Modern ES6+ patterns
- react.md - Hooks, composition, best practices
- nextjs.md - App Router, Server Components
- python.md - Type hints, PEP 8
- go.md - Error handling, defer patterns
- pnpm-workspaces.md - Monorepo commands
### Example Output
```bash
ao init --auto
# Detects:
# ✓ TypeScript + pnpm workspaces
# ✓ React + Next.js
# ✓ Vitest
# Generates:
agentRules: |
Always run tests before pushing.
Use TypeScript strict mode.
Use ESM modules with .js extensions.
Use React best practices (hooks, composition).
Before pushing: pnpm build && pnpm typecheck && pnpm lint && pnpm test
```
## Benefits
- **5 seconds** instead of 5 minutes
- **Zero config knowledge** required
- **Context-aware rules** tailored to your stack
- **Still customizable** - edit the generated config
## Future: --smart (AI-powered)
Flag added but not yet implemented. Will use Claude Code to:
- Analyze CLAUDE.md, CONTRIBUTING.md
- Read CI/CD config
- Generate custom rules based on project patterns
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: detect repo default branch instead of current branch
Fixes Bugbot issue: "Current branch wrongly suggested as default base branch"
## Problem
detectEnvironment was using `git branch --show-current` to suggest
defaultBranch in the config. If a user ran `ao init` while on a feature
branch like `feat/my-work`, the wizard would suggest that feature branch
as the default, causing agents to branch from the wrong base.
## Solution
Added detectDefaultBranch() function with 3 fallback methods:
1. git symbolic-ref refs/remotes/origin/HEAD (most reliable)
2. GitHub API via gh CLI (if ownerRepo known)
3. Check common branch names: main, master, next, develop
Now EnvironmentInfo tracks both:
- currentBranch: The checked-out branch (for display only)
- defaultBranch: The repo's base branch (for config)
## Testing
Tested on feat/seamless-onboarding branch:
- Current branch: feat/seamless-onboarding (displayed)
- Default branch: main (correctly detected for config)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: prevent duplicate framework detection in Python projects
Fixes Bugbot issue: "Duplicate frameworks when multiple Python config files exist"
## Problem
When both requirements.txt and pyproject.toml exist and mention the same
framework (e.g., FastAPI), the detection loop added it to the frameworks
array twice, causing duplicate rules in the generated config.
## Solution
Added addFramework() helper that checks if framework already exists before
adding to the array. Also prevents pytest from being set multiple times as
testFramework.
## Testing
Verified with test repo containing both files with FastAPI:
- Before: Would add 'fastapi' twice
- After: Only adds 'fastapi' once ✓
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address Bugbot review comments
- Remove redundant conditional in --smart flag (both branches were identical)
- Include templates directory in npm package files
* fix: add existence check for base.md template file
Add existsSync guard before reading base.md to handle missing templates gracefully, consistent with other template file reads.
* fix: use direct tool invocation instead of which command
Replace 'which' with direct tool invocation (tmux -V, gh --version)
for better portability on minimal Linux systems where 'which' may
not be installed.
* fix: address Bugbot review comments
- Simplify gh auth status check to rely on exit code instead of output string
- Remove async from synchronous functions (detectProjectType, generateRulesFromTemplates)
* feat: add setup script for one-command installation
Add scripts/setup.sh that:
- Installs pnpm if not present
- Installs dependencies
- Builds all packages
- Links CLI globally
Updated README with simplified setup instructions using the script.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: correct npm link command in setup script
Remove incorrect -g flag from npm link command. The correct syntax is to cd into the package directory and run npm link without flags.
* fix: address Bugbot review comments on init command
- Validate --smart flag requires --auto (prevents silent ignore)
- Fix path validation to check user-specified path (not CWD)
These fixes address medium and low severity issues found by Cursor Bugbot
in PR #66 review.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* docs: add DirectTerminal troubleshooting and fix setup script
- Add TROUBLESHOOTING.md documenting node-pty posix_spawnp error
- Update setup.sh to rebuild node-pty from source (fixes DirectTerminal)
- Ensures seamless onboarding with working terminal out-of-the-box
Resolves DirectTerminal WebSocket failures from incompatible prebuilt binaries.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: resolve variable scope issue in init command validation
- Move path variable outside if block to fix TypeScript scope error
- Only validate path existence if projectId is provided
- Use inline tilde expansion instead of missing expandHome import
Fixes build error that prevented setup.sh from completing.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: automate node-pty rebuild to eliminate terminal issues
- Add postinstall hook to automatically rebuild node-pty after pnpm install
- Create scripts/rebuild-node-pty.js for automatic rebuild with error handling
- Remove manual node-pty rebuild from setup.sh (now automatic)
This ensures DirectTerminal works correctly on every installation without
manual intervention. Fixes posix_spawnp errors from incompatible prebuilt
binaries across different systems and installations.
Resolves issue where users would encounter blank terminals after setup.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* docs: update TROUBLESHOOTING with automatic node-pty rebuild
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* docs: add comprehensive README with quick start guide
- 3-line magical setup: clone → setup → init → start
- Architecture overview with plugin slots table
- Usage examples and auto-reaction configuration
- Links to detailed docs (SETUP.md, TROUBLESHOOTING.md, examples/)
- Philosophy: push not pull, amplify judgment
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: resolve ESLint errors in rebuild-node-pty script
- Add scripts directory configuration to eslint.config.js
- Configure Node.js globals (console, process) for scripts
- Remove unused error variable from catch block
Fixes lint CI failure.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: warn when auto mode uses placeholder repo value
- Detect when 'owner/repo' placeholder is used in --auto mode
- Show warning: 'Could not detect GitHub repository'
- Update next steps to emphasize editing config when placeholder used
- Prevents silent failures when spawning agents with invalid repo
Addresses Bugbot review comment about silent placeholder values.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: wire up live activity detection for agent sessions
Previously, all sessions showed as "idle" on the dashboard because
activity detection was never called. The Agent plugin interface has
detectActivity(terminalOutput) and Runtime has getOutput(), but the
session manager's list() and get() methods never invoked them.
Changes:
- Updated list() to call runtime.getOutput() and agent.detectActivity()
after checking if runtime is alive
- Updated get() with the same activity detection logic
- Captures last 30 lines of terminal output for classification
- Falls back to "idle" if output capture fails (graceful degradation)
- Agent plugins classify output into: active, idle, waiting_input,
blocked, or exited
Testing:
- Added test for activity detection in list() with mocked output
- Added test for activity detection in get() with mocked output
- Added test for graceful fallback when getOutput() fails
- Fixed detectActivity mock (sync not async)
- All 24 tests pass
The dashboard now accurately shows whether agents are actively working,
idle at prompt, waiting for user input, blocked, or exited.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* refactor: extract enrichSessionWithRuntimeState helper
Addresses review comment about duplicated activity detection logic
between list() and get() methods.
Changes:
- Extract activity detection into enrichSessionWithRuntimeState helper
- Both list() and get() now call this shared helper
- Eliminates duplication and makes future changes easier
- All tests still pass (24/24)
Benefits:
- Single source of truth for runtime state enrichment
- Future changes (caching, param tweaks) only need one update
- More maintainable and less error-prone
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* refactor: replace terminal parsing with agent-native activity detection
Replaces hacky terminal output parsing (detectActivity) with deterministic
agent-native state tracking (getActivityState). Each agent now uses its own
internal mechanisms (JSONL files, SQLite, etc.) for reliable activity detection.
## Changes
### Core (`packages/core/src/types.ts`)
- Add `getActivityState(session)` method to Agent interface
- Deprecate `detectActivity(terminalOutput)` with @deprecated tag
### Session Manager (`packages/core/src/session-manager.ts`)
- Update `enrichSessionWithRuntimeState()` to call `agent.getActivityState()`
- Remove terminal output capture and parsing
- Cleaner, more maintainable code
### Claude Code Plugin (`packages/plugins/agent-claude-code/`)
- Implement `getActivityState()` using existing JSONL infrastructure
- Read last JSONL entry and classify by event type:
- user/tool_use → active
- assistant/system → idle
- permission_request → waiting_input
- error → blocked
- stale (>30s) → idle
- Export `toClaudeProjectPath()` for testing
- Add unit tests for path encoding
### Other Agent Plugins (Codex, Aider, OpenCode)
- Add stub `getActivityState()` implementations
- Fall back to process running check
- TODO comments for full implementation using:
- Codex: JSONL rollout files
- Aider: Chat history mtime
- OpenCode: SQLite database
### Tests (`packages/core/src/__tests__/session-manager.test.ts`)
- Update tests to expect `getActivityState()` calls
- Remove tests for deprecated terminal parsing
- All 24 tests pass ✅
## Benefits
✅ **Deterministic** - File-based, not terminal text parsing
✅ **Fast** - Single file read, no regex
✅ **Reliable** - Agent-provided state
✅ **Testable** - Mock files easily
✅ **Maintainable** - Pin versions, test format changes
## Migration Path
- `detectActivity()` is deprecated but not removed (backwards compat)
- Claude Code uses new method immediately
- Other agents use fallback until fully implemented
- Future PR will remove deprecated method
## Research
Agent-native mechanisms documented in:
- OpenAI Codex: JSONL rollout files at ~/.codex/sessions/
- Aider: Chat history at .aider.chat.history.md
- OpenCode: SQLite at ~/.local/share/opencode/opencode.db
See /tmp/activity-detection-redesign.md for full design.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: implement native activity detection for all agent plugins
Implements full getActivityState() for Codex, Aider, and OpenCode:
- **Codex**: Checks JSONL rollout files at ~/.codex/sessions/YYYY/MM/DD/rollout-*.jsonl
Maps event types (user_message, tool_use, approval_request, error) to activity states
- **Aider**: Checks chat history file mtime at .aider.chat.history.md and recent git commits
Detects activity based on file modification and auto-commit behavior
- **OpenCode**: Checks SQLite database mtime at ~/.local/share/opencode/opencode.db
Considers active if database was modified within 30 seconds
All implementations follow the deterministic approach established for Claude Code,
avoiding hacky terminal output parsing in favor of agent-native mechanisms.
Also fixes test mocks to include new getActivityState() method.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address bugbot review comments for activity detection
Fixes three issues identified by Cursor Bugbot:
1. **Claude Code returns "exited" when runtime is alive** (Medium)
- Added process-alive check at start of getActivityState()
- Changed fallback returns from "exited" to "active" when process is running
- Now matches pattern used by other agents (Codex, Aider, OpenCode)
2. **Claude Code missing process-alive check** (Medium)
- Added isProcessRunning() check before file-based detection
- Prevents reporting stale file-based activity when process has exited
- Ensures "exited" is only returned when process is actually dead
3. **Codex reads entire JSONL file into memory** (Medium)
- Replaced readFile() with seeked file handle read
- Now reads only last 4KB (TAIL_READ_BYTES) instead of entire file
- Matches efficient approach used by Claude Code plugin
- Prevents memory/latency issues on long-running sessions
All tests passing (85 for Claude Code, 28 for Codex).
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address three additional bugbot review comments
Fixes three issues identified in the second Cursor Bugbot scan:
1. **Catch block doesn't set activity to idle** (Low)
- Added explicit `session.activity = "idle"` in catch block
- Previously relied on implicit default from metadataToSession
- Now explicitly documents and enforces the fallback behavior
2. **Codex rollout file matching uses wrong session ID** (Medium)
- Removed session ID filtering from findLatestRolloutFile()
- Codex uses internal UUIDs unrelated to orchestrator session IDs
- Now finds most recent rollout file across all dates
- Fixes non-functional activity detection for Codex
3. **Duplicate readLastJsonlEntry across packages** (Low)
- Extracted shared JSONL tail-reading logic to @composio/ao-core/utils
- Removed duplicate implementations from agent-claude-code and agent-codex
- Both plugins now import readLastJsonlEntry and TAIL_READ_BYTES from core
- Net reduction of 21 lines while improving maintainability
All 240 tests passing (127 core + 85 Claude Code + 28 Codex).
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: remove unused TAIL_READ_BYTES import from claude-code plugin
* fix: address third round of bugbot comments - document agent limitations
Fixes three issues from the latest Cursor Bugbot scan:
1. **Codex reads global state not per-session** (High Severity)
- Removed file-based activity detection from Codex plugin
- Codex stores rollout files globally without workspace scoping
- When multiple sessions run, cannot reliably match files to sessions
- Now falls back to process-running check only (conservative)
- Documented limitation with TODO for when Codex adds per-workspace support
2. **OpenCode uses global shared database** (Medium Severity)
- Removed database mtime checking from OpenCode plugin
- OpenCode uses single global SQLite database for all sessions
- Multiple sessions cause activity state bleed
- Now falls back to process-running check only (conservative)
- Documented limitation with TODO for when OpenCode adds per-workspace support
3. **Unused TAIL_READ_BYTES export** (Low Severity)
- Removed TAIL_READ_BYTES from core package exports
- Only used internally by readLastJsonlEntry in utils.ts
- Reduces unnecessary public API surface
Net reduction: 108 lines of code that couldn't work correctly with multiple sessions.
All tests passing (28 Codex + 27 OpenCode).
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* test: rewrite agent integration tests to use getActivityState() instead of detectActivity()
Integration tests were using the deprecated detectActivity() method, which
meant they weren't actually testing the new getActivityState() functionality
that session-manager uses in production.
Changes:
- All agent integration tests now call getActivityState() instead of detectActivity()
- Tests verify getActivityState() returns valid states while running
- Tests verify getActivityState() returns "exited" after process terminates
- Grouped multiple assertions per test run for efficiency
- Updated comments to reflect new behavior (Codex/OpenCode conservative fallback)
- Allow getSessionInfo() to return null for path encoding mismatches
All integration tests pass:
- agent-claude-code: 6 passed (full JSONL-based activity detection)
- agent-codex: 5 passed (conservative fallback)
- agent-opencode: 5 passed (conservative fallback)
- agent-aider: updated but not tested (aider binary not available in CI)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: remove unused imports from integration tests
Fixed ESLint errors:
- Removed unused 'capturePane' imports from all agent integration tests
- Removed unused 'err' variable from catch block
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: return active instead of idle when JSONL entry is null
When readLastJsonlEntry returns null (empty file or read error), getActivityState
now returns "active" instead of "idle" as a conservative fallback. This makes it
consistent with:
- Other fallback cases in the same function (no workspace path, no session file)
- Other agent plugins (Codex, OpenCode, Aider) which return "active" on detection failure
- The stated intent in PR discussion to assume active when process is running
Fixes bugbot comment: https://github.com/ComposioHQ/agent-orchestrator/pull/45#discussion_r2810110669
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: make TAIL_READ_BYTES internal constant instead of exported
TAIL_READ_BYTES is only used internally within readLastJsonlEntry()
and is not exported from the package's index.ts. Removed the export
keyword to make it clear it's an internal implementation detail.
Addresses bugbot comment about unused export.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: implement ao start command for unified orchestrator startup
Adds `ao start` and `ao stop` commands to unify orchestrator and
dashboard startup. Key features:
- Generates CLAUDE.orchestrator.md with project-specific context
- Auto-imports orchestrator prompt via CLAUDE.local.md
- Creates orchestrator tmux session with agent
- Starts Next.js dashboard server
- Supports --no-dashboard, --no-orchestrator, --regenerate flags
- Idempotent operation (safe to run multiple times)
- Computes orchestrator ID from config (not session search)
- Dashboard button always visible for orchestrator terminal
Components:
- packages/core/src/orchestrator-prompt.ts: Prompt generator
- packages/cli/src/commands/start.ts: Start/stop commands
- Modified exports and dashboard UI for orchestrator support
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add automatic metadata updates via Claude Code hooks
CRITICAL: This makes the dashboard work by auto-updating metadata when
agents run git/gh commands. Without this, PRs created by agents never
appear on the dashboard.
Changes:
- packages/core/src/claude-hooks.ts: Setup Claude hooks (settings.json + metadata-updater.sh)
- ao start: Automatically configures Claude hooks in project directory
- ao spawn: Sets AO_SESSION and AO_DATA_DIR env vars for hook script
- metadata-updater.sh: Detects gh pr create, git checkout -b, gh pr merge
How it works:
1. PostToolUse hook fires after every Bash command
2. metadata-updater.sh receives JSON with command and output
3. Pattern matches git/gh commands and updates flat metadata files
4. Dashboard reads metadata files to show PR/branch/status
The .claude directory is symlinked from main repo to worktrees so all
sessions share the same hook config.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* refactor: move hook setup to Agent plugin interface
ARCHITECTURE: Hooks setup must go through the Agent plugin interface,
not be hardcoded for Claude Code. This allows other agents (Codex,
Aider, OpenCode) to implement their own metadata update mechanisms.
Changes:
- Added Agent.setupWorkspaceHooks() method to types.ts
- Added WorkspaceHooksConfig interface
- Implemented setupWorkspaceHooks() in Claude Code plugin
- ao start: calls agent.setupWorkspaceHooks() instead of direct setup
- ao spawn: calls agent.setupWorkspaceHooks() for new worktrees
- Uses $CLAUDE_PROJECT_DIR variable for hook path (works with symlinked .claude)
Each agent plugin now implements its own hook mechanism:
- Claude Code: .claude/settings.json with PostToolUse hook
- Future: Codex, Aider, OpenCode with their own config formats
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address Cursor Bugbot review comments
Fixes 3 issues identified by Cursor Bugbot:
1. HIGH: ao start is now truly idempotent - if orchestrator session exists,
it skips creating the session but still proceeds with dashboard startup
and hook configuration. This allows `ao start` to recover from dashboard
crashes without failing.
2. MEDIUM: Dashboard orchestrator button now finds the actual running
orchestrator session instead of always using the first project. Fallback
to first project ID if no orchestrator is running.
3. LOW: Deduplicated findWebDir() function by moving it to shared utility
lib/web-dir.ts. Now used by both dashboard.ts and start.ts.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: resolve ESLint errors
Fixes 4 ESLint errors identified in CI:
- Added { cause: err } to Error constructors (preserve-caught-error rule)
- Prefixed unused parameter 'config' with underscore in setupWorkspaceHooks
- Prefixed unused variable 'projectId' with underscore in stop command
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address bugbot review comments - markdown escaping and unused module
- Fix markdown code fence escaping in orchestrator-prompt.ts (change
`\\\`` to `\`` for proper markdown rendering)
- Remove unused claude-hooks.ts module (functionality moved to plugin)
- Clean up exports from core/src/index.ts
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address remaining bugbot issues - metadata path, duplication, summary
HIGH severity - Fix metadata path mismatch for worker sessions:
- Add AO_PROJECT_ID environment variable in spawn.ts
- Update metadata-updater hook script to construct correct path:
* Worker sessions: $AO_DATA_DIR/${AO_PROJECT_ID}-sessions/$AO_SESSION
* Orchestrator: $AO_DATA_DIR/$AO_SESSION (no project ID)
- Fixes silent hook failures where PRs/branches never appeared on dashboard
LOW severity - Remove code duplication in hook setup:
- Extract setupHookInWorkspace() helper function (90 lines)
- Refactor setupWorkspaceHooks() to use helper (from 80 lines to 4)
- Refactor postLaunchSetup() to use helper (from 82 lines to 6)
- Eliminates risk of methods drifting out of sync
LOW severity - Fix misleading summary output:
- Change "Orchestrator started" to "Startup complete" when components skipped
- Only show dashboard URL when --no-dashboard NOT used
- Only show session info when --no-orchestrator NOT used
- Show "already running" status when orchestrator exists
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address new bugbot issues - hook quoting, orchestrator link, pkill scope
HIGH severity - Fix Claude hook command quoting:
- Remove embedded double quotes from $CLAUDE_PROJECT_DIR path
- Change from '"$CLAUDE_PROJECT_DIR"/.claude/...' to '$CLAUDE_PROJECT_DIR/.claude/...'
- Prevents hook execution failures if runner treats command as literal path
MEDIUM severity - Fix orchestrator link pointing to nowhere:
- Only show "orchestrator terminal" link when session actually exists
- Remove fallback to computed/hardcoded "ao-orchestrator" ID
- Prevents 404s when user clicks link before starting orchestrator
MEDIUM severity - Fix stop command killing unrelated processes:
- Replace broad `pkill -f "next dev -p ${port}"` with targeted approach
- Use `lsof -ti :${port}` to find exact PID listening on port
- Only kill the specific process, not any process mentioning "next dev"
- Prevents accidentally killing unrelated Next.js dev servers
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: orchestrator metadata path and multi-pid dashboard stop
HIGH severity - Fix orchestrator AO_PROJECT_ID pollution:
- Orchestrator intentionally omits AO_PROJECT_ID (uses flat metadata path)
- agent.getEnvironment() adds AO_PROJECT_ID=project.name
- Object.assign() merged this in, breaking metadata hook path lookup
- Fix: delete environment.AO_PROJECT_ID after merge
- Ensures orchestrator metadata updates work correctly
MEDIUM severity - Fix stopDashboard with multiple PIDs:
- lsof -ti :PORT returns multiple PIDs (one per line) for parent+children
- Passing entire multi-line string to kill fails (can't parse newlines)
- Fix: split stdout by newlines, filter empty, pass PIDs as separate args
- Now correctly stops dashboard even when multiple Node processes exist
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: remove AO_PROJECT_ID from agent plugins to fix metadata path mismatch
The agent.getEnvironment() method was setting AO_PROJECT_ID to
config.projectConfig.name, but different callers have different metadata
path schemes:
- spawn.ts writes to project-specific directories (dataDir/{projectId}-sessions/)
- start.ts writes to flat directories for orchestrator (dataDir/)
- session-manager writes to flat directories (dataDir/)
Setting AO_PROJECT_ID in getEnvironment() caused the metadata updater hook
to look for files in the wrong location for orchestrator and session-manager
flows, breaking automatic metadata updates.
Fix: Remove AO_PROJECT_ID from all agent plugins' getEnvironment() methods
and make it the caller's responsibility to set when using project-specific
directories. Only spawn.ts sets it now.
Changes:
- Remove AO_PROJECT_ID assignment from getEnvironment() in all 4 agent plugins
(claude-code, aider, codex, opencode)
- Update corresponding tests to expect AO_PROJECT_ID to be undefined
- Remove delete environment.AO_PROJECT_ID statement from start.ts (no longer needed)
- Add comments explaining the metadata path scheme contract
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: only run orchestrator setup when actually starting orchestrator
The orchestrator-specific setup steps (generating CLAUDE.orchestrator.md,
configuring CLAUDE.local.md, and setting up agent hooks) were executing
unconditionally, even when --no-orchestrator was passed. This caused
`ao start --no-dashboard` to fail if hook setup had errors, because the
hook setup was fatal and blocked the dashboard from starting.
Fix: Move all orchestrator setup steps inside the `if (opts?.orchestrator !== false)`
guard, and specifically inside the `else` branch (when session doesn't already exist).
Now these steps only run when we're actually creating a new orchestrator session.
This allows:
- `ao start --no-orchestrator` to start only the dashboard
- Skipping setup when orchestrator session already exists
- Setup to be non-blocking for dashboard-only mode
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: eliminate redundant getAgent call by hoisting agent declaration
The agent instance was being created twice in the orchestrator setup block:
- Once at line 237 inside the hook setup try block
- Again at line 253 for getting the launch command
This creates duplicate agent instances unnecessarily. Fixed by declaring
the agent variable before the hook setup try block, allowing it to be
reused for both hook setup and launch command generation.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: address resource leaks and undefined env variable
Fixed 4 Bugbot issues:
1. HIGH: Undefined $CLAUDE_PROJECT_DIR in hook script path
- Changed setupWorkspaceHooks to use absolute path instead of undefined
env variable
- Matches approach used in postLaunchSetup for consistency
2. HIGH: Orchestrator session leaks when metadata write fails
- Added try-catch around tmux session launch and metadata write
- Kills tmux session if metadata write or agent launch fails
- Prevents orphaned sessions from consuming resources
3. MEDIUM: Dashboard process leaks when orchestrator setup fails
- Wrapped orchestrator setup in try-catch block
- Kills dashboard child process if orchestrator setup fails
- Prevents orphaned Next.js server from blocking future starts
4. LOW: Inconsistent indentation (fixed as side effect of restructuring)
- Refactored error handling simplified indentation structure
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add npm publishing support with @composio scope
Set up Changesets for version management, add publish metadata to all 20
packages under the @composio scope, create an unscoped wrapper package
(@composio/agent-orchestrator) for global install, and add a GitHub
Actions release workflow.
- Rename all packages from @agent-orchestrator/* to @composio/ao-*
- Add @composio/agent-orchestrator wrapper (bin shim → @composio/ao-cli)
- Add license, repository, homepage, bugs, files, engines to all packages
- Add .npmrc (access=public), MIT LICENSE file
- Add .changeset/ config with linked versioning for all packages
- Add .github/workflows/release.yml (changesets publish CI)
- Add changeset, version-packages, release scripts to root
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: exclude private web package from release build
The release script now filters out @composio/ao-web, matching the
workflow's existing exclusion and preventing a Next.js build failure
from blocking npm publishing.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: implement CLI with all commands (init, status, spawn, session, send, review-check, dashboard, open)
Implement the full `ao` CLI using Commander.js with 9 commands:
- ao init: Interactive setup wizard that creates agent-orchestrator.yaml
- ao status: Colored terminal table of all sessions (branch, activity, PR, CI)
- ao spawn <project> [issue]: Spawn single agent session (worktree + tmux + agent)
- ao batch-spawn <project> <issues...>: Batch spawn with duplicate detection
- ao session ls|kill|cleanup: Session management subcommands
- ao send <session> <message>: Smart message delivery with busy detection/retry
- ao review-check [project]: Scan PRs for review comments, trigger agent fixes
- ao dashboard: Start the Next.js web dashboard
- ao open [target]: Open session(s) in terminal tabs
Shared libraries:
- lib/shell.ts: Shell command helpers (tmux, git, gh wrappers)
- lib/metadata.ts: Flat metadata file read/write (backwards-compatible format)
- lib/format.ts: Terminal formatting (colors, tables, banners)
All commands code against core interfaces and flat metadata files,
matching the behavior of the reference bash scripts in scripts/.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address PR review — extract shared helpers, fix bugs, add tests
- Extract getTmuxSessions/getTmuxActivity to lib/shell.ts (was duplicated in 5 files)
- Fix readMetadata unsafe cast: return Partial<SessionMetadata> instead
- Fix dashboard webDir path: resolve from package root, not dist/
- Fix dashboard spawn error handling: add child.on("error")
- Fix getNextSessionNumber: escape regex metacharacters in prefix
- Fix fallback worktree creation: use existing branch name, not detached HEAD
- Fix post-create hooks: run before agent launch, not after
- Fix open --new-window: pass option through to openInTerminal
- Fix cleanup dry-run: separate summary message for dry-run mode
- Add Claude session introspection to status command (TTY→PID→CWD→JSONL)
- Add vitest test suite: 72 tests covering commands and lib utilities
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address all PR review issues, add lint/format compliance
Review fixes:
- Extract getTmuxSessions/getTmuxActivity to lib/shell.ts (was duped in 5 files)
- Fix readMetadata: return Partial<SessionMetadata> instead of unsafe cast
- Fix dashboard webDir: use createRequire + __dirname for ESM compat
- Fix dashboard spawn: add child.on("error") handler
- Fix getNextSessionNumber: escape regex metacharacters in prefix
- Fix fallback worktree: use existing branch name, not detached HEAD
- Fix post-create hooks: run before agent launch, not after
- Fix open --new-window: pass option through to openInTerminal
- Fix cleanup dry-run: separate summary message for dry-run mode
- Add Claude session introspection to status (TTY→PID→CWD→JSONL)
- Add port validation in init wizard
- Add clarifying comment for tmux C-u send-keys
Lint/format:
- Add no-console override for CLI package in eslint config
- Fix duplicate imports (node:fs, @agent-orchestrator/core)
- Fix unused variable in send test
- Apply prettier formatting across all files
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address codex review — fd leak, input validation, security hardening
- Wrap openSync/readSync in try/finally to prevent fd leak on error
- Add timeout NaN/<=0 guard in send command
- Add port validation (1-65535) in dashboard command
- Sanitize issueId for git-ref-safe branch names
- Add --project validation in session ls/cleanup
- Fix batch-spawn same-run duplicate detection
- Replace shell interpolation with safe ps + JS filtering
- Wrap temp file usage in try/finally for cleanup on error
- Read only tail of JSONL files to avoid large file loads
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address remaining PR review comments
- Use `=== null` instead of falsy check for git worktree result (spawn.ts)
- Wrap spinner in try/catch so it stops on error (spawn.ts)
- Use exact match instead of substring for issue duplicate detection (metadata.ts)
- Check git worktree remove result before printing success (session.ts)
- Skip banner/headers/footer when --json is passed (status.ts)
- Add error handler on browser-open spawn (dashboard.ts)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address bugbot review — interrupt before send, error resilience, worktree validation
- review-check: send C-c + C-u to interrupt busy agent and clear input before
delivering fix prompt, matching the reference bash script behavior
- review-check: wrap per-session send in try/catch so one failure doesn't abort
remaining sessions
- spawn: check worktree creation return value and throw on failure instead of
continuing with non-existent worktree path
- Update test mock for detached worktree to return success
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address 4 bugbot findings — status type, TTY match, retry safety, comment counting
- spawn: write status "spawning" (not "starting") to match SessionStatus type
- status: use column-based exact TTY match to prevent pts/1 matching pts/10
- send: use null-safe tmux() helper in retry loop instead of throwing exec()
- review-check: use GraphQL reviewThreads to count only unresolved comments,
preventing fix prompts on PRs with all threads resolved
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: GraphQL variable passing and env var name sanitization
- review-check: use -F flags for GraphQL variables instead of string
interpolation to prevent injection via repo config values
- spawn: sanitize env var name by replacing non-alphanumeric chars with
underscores (my-app → MY_APP_SESSION, not MY-APP_SESSION)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add -l flag to tmux send-keys and remove dead activityIndicator code
- Add literal (-l) flag to send-keys so user text isn't interpreted as
tmux key names (e.g. "Enter", "Escape")
- Remove unused activityIndicator function from format.ts and its tests
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: move agent-specific logic from CLI into agent plugins
detectActivity, introspect, and getLaunchCommand now live in the agent
plugins (claude-code, codex, aider) instead of being hardcoded in the
CLI commands. The Agent interface's detectActivity takes a plain string
of terminal output instead of a Session object, making plugins trivially
unit-testable.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: use -f for string GraphQL vars and -l flag for tmux send-keys
Addresses two bugbot findings in review-check.ts:
1. Use -f (string) instead of -F (typed) for owner/name GraphQL
variables to prevent numeric coercion of repo names.
2. Use -l (literal) flag with send-keys and separate Enter call,
matching the established tmux safety pattern in send.ts.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: guard empty terminal output in lifecycle-manager and deduplicate send.ts
- lifecycle-manager: skip detectActivity when terminal output is empty
to prevent false "killed" state when runtime probe returns no data
- send.ts: collapse identical isBusy/isProcessing into single isActive
- tests: use non-empty mock terminal output so detectActivity is exercised
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: validate message before side effects, remove dead updateMetadataField
- send.ts: move message validation before idle-wait loop and C-u clear
so running `ao send session` with no message exits immediately without
touching the tmux session
- metadata.ts: remove exported updateMetadataField (unused in production)
- metadata.test.ts: remove corresponding tests
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add -l flag and separate Enter for spawn prompt send-keys
Use literal flag (-l) when sending the initial prompt via tmux
send-keys in spawn.ts, and send Enter as a separate call. This
matches the pattern used in send.ts and review-check.ts and prevents
tmux from interpreting special characters in the issue ID.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: strict session prefix matching and validate project before use
- session-utils: add matchesPrefix() using regex ^prefix-\d+$ to prevent
cross-project misattribution with overlapping prefixes (e.g. "app" vs
"app-v2")
- Replace all startsWith(`${prefix}-`) calls across status.ts,
review-check.ts, session.ts, and open.ts with matchesPrefix()
- status.ts, review-check.ts: move project validation before the
projects map construction so invalid project IDs exit before any
access to undefined config
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: deduplicate escapeRegex and add -l flag to spawn send-keys
- Export escapeRegex from session-utils.ts, remove duplicate from spawn.ts
- Add -l (literal) flag to tmux send-keys for postCreate hooks and
launch command, with Enter sent separately
- Prevents tmux from interpreting key names in user config or agent
commands
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: update detectActivity to sync string signature across all plugins
After rebase on main, all agent plugins and their tests still had the
old async detectActivity(session: Session) signature. Updated to the
new sync detectActivity(terminalOutput: string): ActivityState across:
- agent-claude-code, agent-codex, agent-aider, agent-opencode plugins
- All plugin unit tests (test terminal output patterns, not process state)
- All integration tests (capture tmux pane output before calling)
- status.ts: use getSessionInfo() instead of nonexistent introspect()
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: integration tests expect idle (not exited) from detectActivity after process exit
detectActivity is a pure terminal-text classifier that returns "idle" for
empty/shell-prompt output. Process exit detection is handled by isProcessRunning.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: check last-line prompt before historical activity patterns in detectActivity
Reorder classifyTerminalOutput to check the last line for a prompt
character before scanning the full buffer for activity indicators like
"Reading" or "Thinking". This prevents historical output in the capture
buffer from causing false "active" results when the agent is idle.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: update pnpm-lock.yaml for web dashboard dependencies
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: preserve state on getOutput failure, check permission prompts before activity indicators
- lifecycle-manager: remove .catch(() => "") from getOutput so failures
reach the outer catch block that preserves stuck/needs_input states
- classifyTerminalOutput: check waiting_input prompts against bottom of
buffer before full-buffer active indicators, preventing historical
"Thinking"/"Reading" text from overriding current permission prompts
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: detect agent exit via isProcessRunning, remove redundant active checks
- lifecycle-manager: replace dead "exited"/"blocked" branches with
isProcessRunning check when detectActivity returns "idle" — correctly
detects agent process exit inside a still-alive tmux session
- classifyTerminalOutput: remove redundant explicit active-indicator
checks that returned the same "active" as the unconditional default
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: wrap killSession in try-catch so cleanup loop continues on failure
One session's kill failure (e.g. archiveMetadata fs error) no longer
aborts the entire cleanup loop, matching the pattern used by batch-spawn
and review-check.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* test: add comprehensive unit and integration tests for CLI
Unit tests (6 new files, 61 tests):
- session-utils: escapeRegex, matchesPrefix, findProjectForSession
- plugins: getAgent, getAgentByName lookup and error cases
- shell: exec, execSilent, tmux, git, gh, getTmuxSessions, getTmuxActivity
- review-check: GraphQL review checking, dry-run, fix prompt delivery
- open: target resolution (all/project/session), fallback, --new-window
- init: rejects existing config file
Integration tests (2 new files, 8 tests):
- spawn-send-kill: real tmux session create, send-keys, kill, graceful re-kill
- session-ls: multi-session listing, per-session capture, activity timestamps
CLI vitest config updated with explicit thread pool parallelization.
Total CLI tests: 68 → 129. Total integration tests: 20 → 28.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address 5 bugbot findings — symlink type, timeout cleanup, file validation, repo format
- spawn.ts: pass `type` param to symlinkSync based on lstat (dir vs file)
- spawn.ts: document TOCTOU gap in getNextSessionNumber (tmux rejects dupes)
- dashboard.ts: store browser timeout and clear it on child exit
- send.ts: wrap file read in try-catch with user-friendly error
- review-check.ts: validate owner/name split before GraphQL query
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat: implement agent plugins (claude-code, codex, aider)
Implement the Agent interface from @agent-orchestrator/core for three
AI coding tools, enabling the orchestrator to launch, detect activity,
and introspect agent sessions.
Claude Code plugin (full implementation):
- JSONL session introspection (summary, session ID, cost, last message type)
- Activity detection via terminal output pattern matching
- Process detection via tmux pane TTY → ps lookup chain
- Launch command generation with permissions/model/prompt support
Codex and Aider plugins (functional stubs):
- Launch command generation with tool-specific flags
- Process detection via TTY lookup
- Stub introspection (no JSONL support yet)
Closes INT-1329
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address PR review — shell escaping, activity detection, tests
- Replace JSON.stringify with POSIX shell escaping (single quotes) for
all agent launch commands to prevent command injection
- Remove `unset CLAUDECODE &&` shell syntax from getLaunchCommand (breaks
execFile); moved to getEnvironment as empty string
- Fix detectActivity returning "exited" for empty pane when process is
confirmed alive — now returns "idle"
- Tighten BLOCKED_PATTERNS to specific error framing (^Error:, ENOENT:,
APIError:, etc.) to avoid false positives on code output
- Check blocked patterns before input patterns to prevent "EACCES:
permission denied" matching INPUT_PATTERNS on the word "permission"
- Remove pgrep fallback in findClaudeProcess (could match wrong session)
- Fix extractCost double-counting: prefer costUSD, fall back to
estimatedCostUsd only when costUSD is absent
- Trim ACTIVE_PATTERNS to stable indicators only
- Add comprehensive vitest tests (127 tests across 3 plugins)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address Codex review + lint fixes
- Use `ps -eo pid,tty,args` instead of `comm` for process detection so
CLI wrappers (node, python) are correctly identified
- Coerce PID from handle.data with Number() + isFinite() to handle
string-serialized PIDs
- Fix token double-counting: prefer `usage` object, only fall back to
flat `inputTokens`/`outputTokens` when `usage` is absent
- Add `--` before Codex prompt to prevent prompts starting with `-`
from being parsed as CLI flags
- Check blocked patterns before input patterns so "EACCES: permission
denied" isn't misclassified as waiting_input
- Fix lint: merge duplicate imports, strict equality, remove non-null
assertions
- Format with prettier
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: multi-pane tmux, EPERM handling, strict process matching
Codex review iteration 1 fixes:
- Iterate all tmux pane TTYs instead of only the first — prevents
false "exited" in multi-pane sessions
- Handle EPERM from process.kill(pid, 0) as "process exists" — EPERM
means the process is alive but we lack permission to signal it
- Use word-boundary regex for process name matching — prevents false
positives on similar names like "claude-code" or paths containing
the substring
- Add tests for EPERM handling, multi-pane detection, and strict
name matching (134 tests total)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: exclude next-env.d.ts from eslint
Auto-generated Next.js type file triggers triple-slash-reference rule.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address Bugbot review — JSON.parse guard, detectActivity pattern priority
- Guard parseJsonlFile against non-object JSON values (null, arrays,
primitives) that would cause TypeError in downstream extractors
- Split null output (non-tmux → "active") from empty output (tmux → "idle")
so non-tmux runtimes don't falsely report idle
- Reorder pattern matching: idle before blocked so stale errors in the
tmux scrollback don't mask an idle prompt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: extract shellEscape to @agent-orchestrator/core
Move the duplicated POSIX shell-escaping utility from all three agent
plugins into packages/core/src/utils.ts and re-export from the package
entry point. Plugins now import { shellEscape } from the shared package.
Addresses Bugbot review comment on PR #5.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add execFileAsync timeouts, use satisfies for plugin exports
- Add { timeout: 30_000 } to all execFileAsync calls (tmux, ps) per
CLAUDE.md convention to prevent hanging on stuck processes
- Replace `const plugin: PluginModule<Agent>` with inline
`export default { ... } satisfies PluginModule<Agent>` to preserve
narrow literal types on manifest fields
Addresses Bugbot review comments on PR #5.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: rename introspect → getSessionInfo, detectActivity uses JSONL
- Rename Agent.introspect() → getSessionInfo(), AgentIntrospection → AgentSessionInfo
- Claude Code detectActivity: replace tmux screen-scraping with JSONL-based
detection (file mtime + last entry type)
- Remove terminal pattern constants (ACTIVE_PATTERNS, IDLE_PATTERNS, etc.)
- Update all tests: JSONL-based activity tests, renamed introspect blocks
- waiting_input/blocked states deferred to hooks-based implementation (#16)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add OpenCode agent plugin, integration tests, and CI workflow
- Add agent-opencode plugin with getLaunchCommand, detectActivity,
isProcessRunning, getSessionInfo (27 unit tests)
- Add integration test suite for all 4 agents (claude-code, codex,
aider, opencode) using real binaries in tmux sessions
- Add GitHub Actions CI workflow for integration tests
- Add test:integration script to root package.json
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: update core test mocks — introspect → getSessionInfo + isProcessing
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address Bugbot review — Windows paths, bytesRead, hardcoded paths
- toClaudeProjectPath: handle Windows backslashes and drive letters
- readLastJsonlEntry: use bytesRead to avoid null character corruption
- Remove duplicate next-env.d.ts eslint ignore entry
- Remove hardcoded developer-specific binary paths from integration tests
- Add isProcessing limitation comments with issue references (#17-19)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Copilot
Madhav Kumar
Dhruv Sharma
harshitsinghbhandari
yyovil
i-trytoohard
Gautam Tayal
Prateek
Harsh Batheja
github-actions[bot]
suraj-markup
Joakim Sigvardt