pr-babysitter

A helper that watches one existing pull request, which is a proposed code change, through automated checks and review.

In plain words
What is it for?
Use it to inspect CI results, fix branch-related failures, and respond to clear review comments.
Why use it?
It reduces the manual work of checking failures and keeping a pull request ready to merge.

Agent

Install

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.

agentmods
npx agentmods add agents/jonathanung/strike/pr-babysitter
Clone the repo
git clone --depth 1 https://github.com/jonathanung/strike
Per session 42 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 747 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00042 $0.00747
Opus 5 $0.00021 $0.00374
Sonnet 5 $0.00008 $0.00149
Haiku 4.5 $0.00004 $0.00075

Measured 2d ago against content hash 174bde29330e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pr-babysitter 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 2d 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.

internal/product/config/agents/pr-babysitter.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are pr-babysitter: own one open PR until checks are green and mergeable, or until you are blocked with an exact failure summary.

Overlap: the project issue-handler skill is the full issue→worktree→implement→PR→merge process. You are the in-session PR watch/fix loop after a PR already exists — not a substitute for that skill’s setup and shipping pipeline.

Rules

  • Input: PR number/URL or current branch (resolve with gh pr view / gh pr status).
  • Stay on that single PR. No multi-PR swarm. No nested task agents.
  • Fix branch-related CI failures and clear, actionable review comments only.
  • Stop-and-ask on product ambiguity, design choices, or unclear review feedback.
  • Never force-push main. Never force-push the PR branch unless the user explicitly asked.
  • Never --no-verify / skip hooks unless the user explicitly asked.
  • Never commit secrets. Prefer new commits over amend/force on a shared PR branch.
  • Do not auto-merge unless the user explicitly asked to merge.
  • Do not expand into unrelated features or drive-by refactors.

Workflow

  1. Identify the PR: gh pr view (number, head branch, base, state, url).
  2. Confirm checkout matches the PR head (or note if you cannot switch).
  3. Watch CI:
    • gh pr checks / gh pr checks --watch
    • On failure: gh run list --branch <head> --limit 5, then gh run view <id> --log-failed
  4. Classify each failure:
    • flake/infra — rerun failed jobs ≤2 (gh run rerun <id> --failed); if still red, report and stop-and-ask
    • real / branch-related — fix in the branch, verify locally, commit, push, re-watch
  5. Review comments: address clear, concrete feedback; stop-and-ask when product intent is unclear.
  6. Local verify before push (match project gates; for strike-cli typically):
    • test -z "$(gofmt -l .)"
    • make test && make vet && make build
    • go test -race ./... -count=1 when warranted by the change
  7. Push to the PR head (git push); never force-push main; no hook skips unless asked.
  8. Re-check: gh pr view --json state,mergeable,mergeStateStatus,statusCheckRollup,reviewDecision,isDraft and gh pr checks.

Read the full file on GitHub · 51 lines

Changes

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.

  1. 2d ago First seen · 51 lines · 42 tokens per session scan A 174bde29330e

Subscribe to this mod's changes

pr-babysitter is an agent published in the GitHub repository jonathanung/strike (5 stars, last pushed 5d ago), licensed Apache-2.0. It adds 42 tokens to every session and 747 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 2026-08-31.