SKILL
Dependency Release Monitor Loop
Plain markdown at a stable URL. Fetch it, follow it.
Use when deploying the loopmaster.ai dependency-release-monitor template to watch releases, assess risk, and route high-risk updates for review.
Dependency release monitor loop skill
Use this pack to install or verify dependency-release-monitor without duplicating the template guide. It reports release evidence and recommendations; it does not merge dependency changes.
Canonical references:
templates/dependency-release-monitor/loop.yaml— permissions, gates, state paths, verification commands, and backend hints.templates/dependency-release-monitor/AGENT-INSTALL.md— complete backend install guide.templates/dependency-release-monitor/scripts/dependency_release_monitor.py— checked-in runner used by every backend.templates/dependency-release-monitor/examples/dry-run-dependencies.json— local-only dependency config.templates/dependency-release-monitor/examples/dependencies.json— editable dependency config for real installs.
If this pack conflicts with templates/dependency-release-monitor/AGENT-INSTALL.md, follow the template guide and fix this pack.
Preflight from repo root
test -f templates/dependency-release-monitor/loop.yaml
test -f templates/dependency-release-monitor/AGENT-INSTALL.md
python3 templates/dependency-release-monitor/scripts/dependency_release_monitor.py --help >/tmp/loopmaster-dependency-release-monitor-help.txt
uv run --with check-jsonschema --with pyyaml check-jsonschema --schemafile spec/loop.schema.json templates/dependency-release-monitor/loop.yamlExpected result: each command exits 0 and manifest validation prints ok -- validation done.
Pick a backend
- cron: shortest local scheduler path; writes report/state from a persistent env file.
- GitHub Actions: repository scheduler; writes report/state artifacts and opens a PR.
- agent-kanban: recurring agent task; the task runs the helper and creates follow-up cards for high-risk or low-confidence releases.
Use templates/dependency-release-monitor/AGENT-INSTALL.md for exact backend setup.
Fast scratch verification
This is not a live install. It proves release detection and duplicate suppression under /tmp.
rm -rf /tmp/loopmaster-dependency-release-monitor /tmp/loopmaster-dependency-release-monitor-state.json
python3 templates/dependency-release-monitor/scripts/dependency_release_monitor.py \
--dependencies templates/dependency-release-monitor/examples/dry-run-dependencies.json \
--state /tmp/loopmaster-dependency-release-monitor-state.json \
--output-dir /tmp/loopmaster-dependency-release-monitor \
--max-releases 5
python3 templates/dependency-release-monitor/scripts/dependency_release_monitor.py \
--dependencies templates/dependency-release-monitor/examples/dry-run-dependencies.json \
--state /tmp/loopmaster-dependency-release-monitor-state.json \
--output-dir /tmp/loopmaster-dependency-release-monitor \
--max-releases 5Expected stdout shape: first run prints DEPENDENCY_RELEASE_REPORT=..., DEPENDENCIES_CHECKED=2, NEW_RELEASES=2, RECOMMEND_ACT=1, and FETCH_ERRORS=0; second run prints NEW_RELEASES=0, RECOMMEND_ACT=0, and FETCH_ERRORS=0.
Stop or roll back
- cron: remove the crontab line that runs
templates/dependency-release-monitor/scripts/dependency_release_monitor.py. - GitHub Actions: disable or delete the copied workflow in a PR.
- agent-kanban: pause or remove the recurring task.
- Bad report: close/revert the report PR or delete local artifacts after recording the failure.
First-week ownership checklist
- Confirm every recommendation cites release evidence or a fetch error.
- Confirm duplicate suppression works on repeated runs.
- Confirm low-confidence or high-risk recommendations become review tasks.
- Keep major upgrades, dependency merges, credential changes, and public publishing human-gated.