Summary
PR #34 bumps two pinned tool versions in the Dockerfile, but the standard CI pipeline did not run on the PR — auto-merge is therefore blocked and this issue tracks the human review.
Bumps
| Tool |
Old |
New |
Source |
Risk |
OpenChamber (npm @openchamber/web) |
1.13.7 |
1.13.8 |
openchamber/openchamber |
Patch, no breaking changes |
GitLab CLI (glab) |
1.103.0 |
1.105.0 |
gitlab-org/cli v1.105.0 |
Minor with glab orbit --format → --response-format rename |
OpenChamber 1.13.8 — release highlights
- Startup hang (~20s) on launch fixed
- When a separate OpenCode is already running, OpenChamber now starts its own server instead of attaching (fixes "could not finish initialization")
- New Follow-up behavior setting (Steer vs Queue) replaces the old queue-mode toggle
- Worktree group deletion now cleans up subagent sessions
- New optional env vars:
OPENCODE_HOST, OPENCODE_PORT, OPENCODE_SKIP_START (additive, opt-in)
glab 1.105.0 — release highlights
- ✨
feat(orbit): rename --format flag to --response-format — breaking for any glab orbit --format=... usage (none in this repo)
- 🐛
fix(orbit): handle new nested status response shape
- 🐛
fix: route deprecation warnings to stderr
- 📦 Dep bump:
gitlab.com/gitlab-org/api/client-go/v2 to v2.40.1
CI status (blocker)
- Workflow run: #232
- Conclusion:
action_required
- Jobs executed: 0 (
build, test, scan did not start)
- Branch protection on
main: none (so the action_required is not a branch-rule approval gate)
- Previous PR (#33) had 3 active jobs, all green — so the standard pipeline works for human PRs
The PR body claims "Integration tests passed on the updated image" and "No new critical Grype findings", but those come from the bot's own pre-PR validation, not from the project's standard CI gate.
Recommended next steps
- Decide policy: is this bot's pre-PR validation considered sufficient to skip the standard CI gate? If yes, document that as the project's trust model and merge. If no, force the standard CI to run.
- Force CI to run (easiest): push a no-op commit, or close + reopen the PR. The
ci.yml trigger (pull_request: branches: [main]) should then start a fresh run with the build/test/scan jobs.
- Re-evaluate glab 1.105.0 after CI passes — the
orbit flag rename is a minor break in semver terms but does not affect this repo's Dockerfile (no glab orbit --format calls).
Files in the PR
Dockerfile — 2 lines
README.md — 1 line
docs/CHANGELOG.md — adds [1.1.7] entry
test-output.txt — version sync
References
Summary
PR #34 bumps two pinned tool versions in the Dockerfile, but the standard CI pipeline did not run on the PR — auto-merge is therefore blocked and this issue tracks the human review.
Bumps
@openchamber/web)glab)glab orbit --format→--response-formatrenameOpenChamber 1.13.8 — release highlights
OPENCODE_HOST,OPENCODE_PORT,OPENCODE_SKIP_START(additive, opt-in)glab 1.105.0 — release highlights
feat(orbit): rename --format flag to --response-format— breaking for anyglab orbit --format=...usage (none in this repo)fix(orbit): handle new nested status response shapefix: route deprecation warnings to stderrgitlab.com/gitlab-org/api/client-go/v2to v2.40.1CI status (blocker)
action_requiredbuild,test,scandid not start)main: none (so theaction_requiredis not a branch-rule approval gate)The PR body claims "Integration tests passed on the updated image" and "No new critical Grype findings", but those come from the bot's own pre-PR validation, not from the project's standard CI gate.
Recommended next steps
ci.ymltrigger (pull_request: branches: [main]) should then start a fresh run with the build/test/scan jobs.orbitflag rename is a minor break in semver terms but does not affect this repo's Dockerfile (noglab orbit --formatcalls).Files in the PR
Dockerfile— 2 linesREADME.md— 1 linedocs/CHANGELOG.md— adds[1.1.7]entrytest-output.txt— version syncReferences