Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ckorhonen/claude-skills --skill babysit-prgit clone --depth 1 https://github.com/ckorhonen/claude-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ckorhonen/claude-skills/babysit-pr)<a href="https://agentmods.dev/skills/ckorhonen/claude-skills/babysit-pr"><img src="https://agentmods.dev/badge/skills/ckorhonen/claude-skills/babysit-pr/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ckorhonen/claude-skills/babysit-pr"><img src="https://agentmods.dev/badge/skills/ckorhonen/claude-skills/babysit-pr.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00048 | $0.02069 |
| Opus 5 | $0.00024 | $0.01035 |
| Sonnet 5 | $0.00010 | $0.00414 |
| Haiku 4.5 | $0.00005 | $0.00207 |
Grade A, and why
babysit-pr scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 196 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PR Lifecycle Operator
Objective
Own the pull request from creation through post-merge validation with as little manual babysitting as possible.
Terminal outcomes:
- The PR is merged and the deployment looks healthy.
- The PR is merge-safe and waiting on humans.
- A blocker requires user help: ambiguous product intent, permissions, unrelated dirty worktree, external-only outage, or missing deployment visibility.
Do not stop at "PR created" or "CI is green once." Keep following the PR until one of the terminal outcomes is true.
Common Failure Modes
These gotchas appear during real PR automation and require proactive handling:
Watcher Script Exits Silently (Template Resolution)
The watcher can exit unexpectedly if PR template resolution fails during preflight. Root cause: resolve_pr_template.py may fail to parse custom template syntax, or template file is unreadable. Mitigation: always run the resolver in --json mode first to validate template before starting the watcher. If resolution fails, fall back to the bundled default template at assets/default_pr_template.md.
Multiple Watchers on Same PR (Duplicate Comments)
Running multiple gh_pr_watch.py instances for the same PR causes duplicate comment processing and can result in duplicate bot comments or conflicting retry actions. This typically happens when the watcher is restarted without properly stopping the previous session. Mitigation: use one watcher session per PR. Verify no orphaned watcher processes exist before starting: ps aux | grep gh_pr_watch. Kill any stale sessions before resuming.
CI Retry Logic Triggers on Non-Flaky Failures (Quota Waste)
The retry logic may consume the CI retry budget on failures that are not actually transient (e.g., legitimate test failures from branch-specific issues). This wastes quota and can block progress on legitimate issues. Mitigation: always inspect CI logs with gh run view <run-id> --log-failed before retrying. Confirm the failure is truly flaky or unrelated to branch changes. Apply the heuristics in references/heuristics.md strictly before calling --retry-failed-now.
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 501 B
- assets/default_pr_template.md 211 B
- references/deployment-monitoring.md 2.2 KB
- references/github-api-notes.md 1.2 KB
- references/heuristics.md 1.3 KB
- scripts/gh_deploy_watch.py 8.8 KB runs code
- scripts/gh_pr_watch.py 28 KB runs code
- scripts/resolve_pr_template.py 2.7 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 196 lines · 48 tokens per session scan A 03a7843b7a5e
babysit-pr is a skill published in the GitHub repository ckorhonen/claude-skills (14 stars, last pushed 2mo ago), licensed MIT. It adds 48 tokens to every session and 2,069 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
create-pull-request
Use when asked to create or update a PR, revise its description, or link issue references to its body. Not for multi-PR stacks: use gate-and-merge. Not for releases: use git-workflow-and-versioning.
gate-and-merge
Use when landing a queue of open PRs: gate each PR, sweep its review feedback to root cause, then merge, repair, hold, or close it. Human-only.
review-pr
Multi-agent PR review with four modes (review, re-review, self-review, address-feedback) - spawns parallel subagents, saves diff to /tmp for context efficiency, supports file exclusion patterns.
pr-review-canvas-html
Use when asked to render a GitHub PR as a standalone review HTML page. Fetches PR data via gh API, renders diffs with move detection, and serves the artifact on a local port. Not for Cursor Canvas output — use pr-review-canvas.
land-contribution
Use when a maintainer or collaborator explicitly asks to review and land one external pull request. Don't use for internal pull requests or landing without preserving contributor authorship.
resolve-pr-feedback
Use when handling GitHub PR review feedback: autonomous fix-and-resolve, interactive per-comment walkthrough, or read-only severity summary. Not for non-GitHub feedback: use resolve.