Commit graph

176 commits

Author SHA1 Message Date
openhands
a9b8144401 fix: bundled dust cleanup — AGENTS.md (#420)
- Add phase-handler.sh to dev Key files (#332)
- Document ci_required_for_pr() in ci-helpers.sh (#347)
- Update planner trigger: weekly → daily, note optional project arg (#416)
- Add dev-agent.sh, action-agent.sh to formula-session.sh sourced-by (#418)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 16:02:29 +00:00
openhands
24cf01028a chore: gardener housekeeping 2026-03-21
- Update AGENTS.md watermarks (all 10 files) to HEAD 038581e5
- Content already current from recent gardener migration and setup PRs

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-21 14:33:26 +00:00
openhands
b630c6fcc1 fix: gardener migration — run-gardener.toml via direct cron, remove legacy scripts (#490)
Rewrite gardener-run.sh as direct cron runner (matching supervisor/planner/
predictor pattern): lock guard, memory check, worktree, tmux session with
Claude sonnet + formulas/run-gardener.toml, phase monitoring, cleanup.

- Delete gardener-poll.sh and gardener-agent.sh (superseded)
- Extract consume_escalation_reply() to lib/formula-session.sh (shared
  by gardener and supervisor, eliminates duplicate blocks)
- Update AGENTS.md, gardener/AGENTS.md, lib/AGENTS.md, CI smoke test,
  and cross-references

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 13:09:17 +00:00
openhands
f480cbe5d0 chore: gardener housekeeping 2026-03-21
Progressive disclosure split of AGENTS.md (487→152 lines):
- Extracted per-directory AGENTS.md files for all 8 agents + lib/
- Root AGENTS.md now serves as a table of contents with summary table
- All watermarks updated to 16e430e

Grooming results:
- Promoted #469 (WATCH flow missing curl) and #436 (idle_pane_count bug) to backlog
- 12 dust items classified, no groups ripe for bundling yet
- No blocked issues, no AD violations
2026-03-21 12:44:23 +00:00
openhands
aebda221f0 fix: file-action-issue.sh dedup check: paginate or raise limit beyond 50 (#383)
Replace codeberg_api GET with codeberg_api_all for the dedup fetch,
which paginates through all open action issues instead of capping at 50.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 11:45:03 +00:00
openhands
38f9b29867 Revert "Merge pull request 'fix: gardener-poll.sh needs to file action issues (not call gardener-agent.sh directly) (#367)' (#474) from fix/issue-367 into main"
This reverts commit cbd52972f3, reversing
changes made to 0dd607c1bb.
2026-03-21 11:00:48 +00:00
openhands
8193e7bc96 fix: extract build_formula_issue_body to eliminate duplicate code blocks
Move TOML frontmatter construction into a shared helper in
lib/file-action-issue.sh, used by both gardener-poll.sh and
gardener-run.sh. Fixes CI duplicate-detection failure.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 10:23:00 +00:00
openhands
1c8c6a8caf fix: add AKIA and 0x private key redaction, preserve scan diagnostics
- redact_secrets now covers AWS AKIA keys and 0x-prefixed private keys,
  matching all patterns that scan_for_secrets detects
- file-action-issue.sh no longer suppresses stderr from scan_for_secrets,
  so blocked-issue diagnostics reach the caller's log

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 10:00:13 +00:00
openhands
0b453fdab0 fix: fix: action formulas must not contain secrets in issue body (#291)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 09:54:21 +00:00
openhands
a2fe3ecb83 fix: formula agents run in isolated git worktrees — no session collisions (#460)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 07:51:27 +00:00
openhands
02f483ff71 fix: matrix_listener.sh drops dev/review injections — PROJECT_NAME is unset (#354)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 07:38:39 +00:00
openhands
60d7920aa1 fix: load-project.sh does not re-derive CODEBERG_WEB on TOML-based repo override (#147)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 06:24:41 +00:00
openhands
d29c6ad1c9 fix: dev-poll.sh: redundant manual state exclusions alongside ci_passed (#113)
Add ci_failed() helper to lib/ci-helpers.sh and replace three compound
`! ci_passed && CI_STATE != "" && != "pending" && != "unknown"` patterns
in dev/dev-poll.sh with the cleaner ci_failed() call.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 06:10:39 +00:00
openhands
bda9240268 refactor: extract ensure_blocked_label_id to lib/ci-helpers.sh (#352)
Move ensure_blocked_label_id() from dev/phase-handler.sh into
lib/ci-helpers.sh to eliminate the duplicate blocked-label creation
curl block that existed in both phase-handler.sh and dev-poll.sh.

Both dev-agent.sh and action-agent.sh now source lib/ci-helpers.sh
so the function is available when phase-handler.sh calls it.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 05:06:12 +00:00
openhands
ab122c9701 fix: PHASE:needs_human missing from crash-path terminal set in monitor_phase_loop (#342)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 03:50:21 +00:00
openhands
e3895ad3ac fix: feat: SessionStart compact hook re-injects phase protocol after context compaction (#274)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 23:27:32 +00:00
openhands
f78fbc1da6 fix: bundled dust cleanup — lib/matrix_listener.sh (#264)
- Remove dead ROOM_ENCODED and EVENT_ID variables from matrix_listener.sh
  (were suppressed with SC2034 instead of removed)
- Remove dead REPO variable from dev-poll.sh and review-poll.sh
- Update header comment in matrix_listener.sh to list all 5 reply-routing
  cases (supervisor, gardener, dev, review, vault, action)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 21:40:31 +00:00
openhands
26d20af48c fix: address review — scratch file survives crash, cap read size, fix instruction (#262)
- Remove SCRATCH_FILE from action-agent cleanup() trap so it survives crashes
- Change instruction to note contents already injected (avoid wasted tool call)
- Cap scratch file read at 8KB via head -c 8192
- Move predictor scratch instruction after formula (consistent placement)
- Remove redundant FINAL_PHASE re-reads in planner/predictor

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 20:58:32 +00:00
openhands
7199bbf9b5 fix: feat: agents flush context to scratch file before compaction (#262)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 20:12:45 +00:00
openhands
3cd047a7e0 fix: P2e and classify_pipeline_failure() use divergent infra heuristics (#251)
Extract shared is_infra_step() in lib/ci-helpers.sh capturing the union of
infra-detection heuristics from both P2e and classify_pipeline_failure():
- Clone/git step exit 128 (connection failure)
- Any step exit 137 (OOM/signal 9)
- Log-pattern matching (timeouts, connection failures)

Update classify_pipeline_failure() to use is_infra_step() with log fetching
and "any infra step" aggregation (matching P2e semantics). Simplify P2e to
delegate to classify_pipeline_failure(). Update P2f caller for new output
format ("infra <reason>").

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 19:19:29 +00:00
openhands
db66e35556 fix: lib/matrix_listener.sh: review case reads a separate /tmp/review-thread-map (col 2) instead of the standard THREAD_MAP (col 4) (#238)
- matrix_listener.sh: review case now reads PR number from column 4 of
  the standard $THREAD_MAP instead of column 2 of /tmp/review-thread-map
- review-pr.sh: pass PR_NUMBER as context_tag (4th arg) to matrix_send
  so the standard MATRIX_THREAD_MAP has it in column 4; remove separate
  /tmp/review-thread-map write
- review-poll.sh: prune from MATRIX_THREAD_MAP instead of the removed
  /tmp/review-thread-map

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 18:21:42 +00:00
openhands
3ea8c183a4 refactor: extract shared prompt footer and monitor loop into formula-session.sh
Eliminates 7 duplicate code blocks between planner-run.sh and predictor-run.sh
flagged by CI duplicate-detection. Adds build_prompt_footer() and
run_formula_and_monitor() helpers to lib/formula-session.sh.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 17:41:52 +00:00
openhands
d05a62bee0 fix: .env contains harb-specific defaults that leak into non-harb sessions (#398)
Remove per-project settings from .env.example and env.sh defaults:
- CODEBERG_REPO no longer defaults to johba/harb
- WOODPECKER_REPO_ID no longer defaults to 2
- Remove deprecated HARB_REPO_ROOT alias
- .env.example now points to projects/*.toml for per-project config

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 16:38:08 +01:00
openhands
5bac4a8409 fix: extract lib/formula-session.sh to eliminate duplicate code blocks
Shared helpers for formula-driven cron agents: lock, memory guard,
formula loading, context building, session startup, crash recovery.

- planner-run.sh uses shared helpers instead of inline code
- gardener-agent.sh delegates crash recovery to formula_phase_callback
- agent-smoke.sh updated for renamed planner script + new lib file

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 13:53:33 +00:00
openhands
59b6d76afa fix: extract file_action_issue helper to eliminate duplicate code blocks
CI duplicate-detection flagged shared action-issue filing pattern between
gardener-run.sh and planner-poll.sh. Extract into lib/file-action-issue.sh
and refactor both scripts to use it.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 12:49:08 +00:00
openhands
898f958196 fix: add model=opus to run-planner formula and wire through action-agent
TOML declares model = "opus". planner-poll.sh includes model: opus in
the issue YAML front matter. action-agent.sh extracts it and exports
CLAUDE_MODEL. create_agent_session passes --model to claude if set.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 09:08:17 +00:00
openhands
6f30614dda fix: fix: guard blocks merge injection — Claude closes issue without merging (#344)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 07:37:32 +00:00
openhands
2a3e5eef90 fix: ci_required_for_pr should paginate /pulls/{pr}/files (#273)
Replace single-page curl fetch with codeberg_api_all() which paginates
through all pages of PR files. This ensures large PRs with 50+ files
correctly detect code files on all pages, preventing CI from being
incorrectly bypassed.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 06:55:00 +00:00
openhands
f66fcd666c fix: address review — terminal phase guard, explicit marker var, test coverage
- Guard against overwriting terminal phases (PHASE:done, PHASE:merged)
  in on-stop-failure.sh to prevent false failures from same-turn race
- Declare sf_phase_marker explicitly in StopFailure block instead of
  relying on phase_marker from PostToolUse block
- Add authentication_failed test (10c) and terminal phase guard tests
  (10g, 10h)
- Fix fragile nested command substitution in test 10f fail() message

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 01:52:46 +00:00
openhands
eaf2841494 fix: feat: StopFailure hook writes phase file on API error / rate limit (#275)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 01:43:00 +00:00
openhands
0a095c4656 fix: feat: SessionEnd hook for guaranteed cleanup on session exit (#276)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 01:11:26 +00:00
openhands
c8b09d1c72 fix: address review — whitelist /tmp, block bare force-push, improve checkout guard
- Guard 2: add /tmp/* to allowlist so normal temp file cleanup is not blocked
- Guard 1: block bare `git push --force` (no branch arg) since upstream may
  point to primary branch
- Guard 4: allow flags between verb and branch (`git switch --detach main`),
  escape branch name for regex safety, exclude -b/-B/-c/-C (branch creation)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 00:16:57 +00:00
openhands
de8dcef81e fix: feat: PreToolUse hook guards destructive operations in dev-agent sessions (#277)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 00:10:27 +00:00
openhands
424a53c9f7 fix: feat: stream action-agent Claude output to Matrix thread (#293)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-19 23:43:29 +00:00
openhands
63d4179f38 fix: Deduplicate hook entries in settings.json on repeated create_agent_session calls (#299)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-19 21:19:58 +00:00
openhands
809dd93c3b fix: distinguish phase file writes from reads in PostToolUse hook
- Parse tool_name via jq: Write tool checks file_path match,
  Bash tool checks for redirect operator (>) with phase file path
- Reads (cat, head) no longer trigger false-positive markers
- Split guard into separate statements for clarity
- Move marker cleanup inside hook-install guard
- Expand tests: 5 cases covering Bash write, Write tool, Bash read,
  unrelated Bash, and Write to different file

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-19 18:14:49 +00:00
openhands
ac04dc29a6 fix: feat: PostToolUse hook detects phase file writes in real-time (eliminates polling latency) (#278)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-19 17:55:06 +00:00
johba
ab3efa2402 fix: replace fragile pane grep with Stop hook for idle detection (#272)
## Summary

- Claude Code v2.1.79 permanently shows `❯` in the input area even while actively thinking, causing `monitor_phase_loop` to false-positive on idle detection and kill working sessions after 90 seconds
- Replace `tmux capture-pane | grep ❯` with a Claude Code Stop hook (`lib/hooks/on-idle-stop.sh`) that writes a marker file only when Claude actually finishes responding
- Hook is installed per-worktree in `.claude/settings.json` by `create_agent_session`; marker cleaned up on inject/kill

## Test plan

- [x] Verified hook installs correctly in fresh worktree
- [x] Verified marker file appears only after Claude finishes responding (not during active thinking)
- [x] Verified live dev-agent session picks up fix and Claude works without being killed
- [x] Verified `agent_inject_into_session` clears marker before new work

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: openhands <openhands@all-hands.dev>
Reviewed-on: https://codeberg.org/johba/disinto/pulls/272
2026-03-19 14:57:54 +01:00
openhands
1ab700c87a fix: feat: review + dev-poll skip CI gate for non-code PRs (#266)
Add diff_has_code_files() and ci_required_for_pr() helpers to
ci-helpers.sh. Non-code PRs (docs/*, formulas/*, evidence/*, *.md)
that have no CI results now skip the CI gate instead of being stuck
forever.

Applied to:
- review-pr.sh: CI gate skipped for non-code PRs
- review-poll.sh: CI gate skipped for non-code PRs
- dev-poll.sh: CI state treated as "success" for non-code PRs in
  orphan, stuck-PR, and backlog merge paths

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-19 13:48:00 +00:00
johba
d5c2c213a3 fix: bug: gardener hangs forever when Claude finishes without writing phase file (#261) (#263)
Fixes #261

## Changes
Fixed gardener hanging forever when Claude skips phase protocol. Three changes: (1) gardener-agent.sh: replaced 999999s timeout with 7200s (2h, matching dev-agent); (2) lib/agent-session.sh: added idle-prompt detection to monitor_phase_loop — if Claude returns to the ❯ prompt for 3 consecutive polls with no phase file written, exits immediately with _MONITOR_LOOP_EXIT=idle_prompt (only fires when phase file is empty, so awaiting_ci/review waits are unaffected); (3) gardener prompt: removed 'no time limit' wording, replaced with explicit phase-write requirement.

Co-authored-by: openhands <openhands@all-hands.dev>
Reviewed-on: https://codeberg.org/johba/disinto/pulls/263
Reviewed-by: Disinto_bot <disinto_bot@noreply.codeberg.org>
2026-03-19 13:47:10 +01:00
openhands
97ffdca95c fix: address review feedback on escalation triage (#185)
- supervisor-poll.sh: check PR state before retrigger; discard stale escalations
  for closed/merged PRs instead of pushing to their branches
- supervisor-poll.sh: bump escalation ts to now on failed retrigger push, so
  the 30-min cooldown resets and alert flooding is avoided on persistent failures
- ci-helpers.sh: require at least one confirmed infra step before returning
  "infra"; prevents false-positive when all step names are empty strings
- ci-helpers.sh: clarify header comment to distinguish per-function requirements
- AGENTS.md: document classify_pipeline_failure() in ci-helpers.sh table row

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-19 09:03:03 +00:00
openhands
051ff39144 fix: feat: supervisor auto-retriggers CI after infra-only exhaustion (#185)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-19 08:51:30 +00:00
openhands
4e56d14e6a fix: feat: action-agent — tmux + Claude + formula for operational tasks (#139)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-19 07:25:25 +00:00
openhands
17907063f6 fix: codeberg_api_all does not accept a custom token parameter (#198)
Add optional second TOKEN parameter to codeberg_api_all in lib/env.sh,
defaulting to $CODEBERG_TOKEN. Pass $REVIEW_BOT_TOKEN at the dismiss
block in review-pr.sh so reviews are fetched as the review bot account.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-19 01:25:44 +00:00
openhands
e853949b47 fix: Callbacks can't see the resolved _session from monitor_phase_loop (#200)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-19 01:05:21 +00:00
openhands
e0f37ede12 fix: agent-session.sh: monitor_phase_loop should accept SESSION_NAME as a parameter (#187)
Add optional 4th parameter to monitor_phase_loop for SESSION_NAME,
falling back to the $SESSION_NAME global for backwards-compatibility.
Document the full function signature in both the file header and inline comment.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-18 20:15:14 +00:00
openhands
61c518116a fix: fix: phase-handler.sh references LAST_PHASE_MTIME which is now internal to monitor_phase_loop (#181)
Export LAST_PHASE_MTIME from monitor_phase_loop before invoking the callback
so that phase-handler.sh can compare phase file mtimes inside the awaiting_review
inner poll loop without hitting an unbound variable error under set -u.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-18 17:26:54 +00:00
openhands
55b5628e24 feat: CI smoke test — syntax check + function resolution for all agent scripts (#177)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-18 17:11:02 +00:00
openhands
014965494b fix: implement monitor_phase_loop in agent-session.sh
Generic phase monitoring loop with callback for agent-specific handling.
Handles: idle timeout, session crash detection, terminal phases (done,
failed, needs_human, merged). Sets _MONITOR_LOOP_EXIT for the caller.
2026-03-18 16:40:51 +00:00
openhands
d83098f382 fix: pass SESSION_NAME to all agent-session.sh function calls
Library functions need explicit session name argument — they no longer
have closure over $SESSION_NAME from the parent script.

- agent_kill_session: add $SESSION_NAME to all 11 call sites
- agent_inject_into_session: add $SESSION_NAME to all call sites in
  phase-handler.sh and gardener-agent.sh
- agent_kill_session: guard against missing arg (defensive)
2026-03-18 16:24:58 +00:00