Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/jscraik/Agent-Skillsnpx agentmods add skills/jscraik/agent-skills/pr-green-sweepWrote 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/jscraik/agent-skills/pr-green-sweep)<a href="https://agentmods.dev/skills/jscraik/agent-skills/pr-green-sweep"><img src="https://agentmods.dev/badge/skills/jscraik/agent-skills/pr-green-sweep/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/jscraik/agent-skills/pr-green-sweep"><img src="https://agentmods.dev/badge/skills/jscraik/agent-skills/pr-green-sweep.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00051 | $0.02532 |
| Opus 5 | $0.00026 | $0.01266 |
| Sonnet 5 | $0.00010 | $0.00506 |
| Haiku 4.5 | $0.00005 | $0.00253 |
Grade A, and why
pr-green-sweep 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 8d 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 — 204 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PR Green Sweep
Own PR closeout from live evidence through merge and cleanup. Use an evidence-backed action queue and one continuation heartbeat; a required external failure blocks its PR while unrelated entries keep their safe action.
When To Use
- The user asks to monitor, fix, or keep rotating through open PRs until green.
- Open PRs need GitHub plugin/gh truth, CodeRabbit review fixes, CircleCI log triage, Context7 docs checks, merge, or cleanup.
- The user wants merged PR branches and worktrees pruned after merge proof.
Gotchas
- Read-only PR summaries when the user asked for until-green follow-through.
- Local test debugging with no PR, review, merge, or continuation workflow.
- Broadening from the current repo unless the user says "all", "everything", "broad", or names multiple repos/orgs.
- Admin merges, force pushes, remote branch deletion, or worktree deletion without explicit approval for that action class.
- Declaring green from local tests while live required checks are pending, failing, stale, or attached to an older head SHA.
Inputs
Use the current repo unless the user names a broader scope. Gather target PRs,
heartbeat cadence, merge/check policy, approval posture, and GitHub/CodeRabbit/
CircleCI auth context. Start with two or three focused failure surfaces before
broadening a multi-repository sweep. Run credentialed CircleCI commands through
the host-configured auth-backed wrapper with ~/.codex/.env; discover that
wrapper from the active Codex environment rather than invoking an unqualified
run-auth-backed.sh or op directly, and never print secrets.
Outputs
For a non-trivial response, emit heartbeat_status first. Then emit
schema_version: 1 with an action queue
(auto_fixable_now, needs_merge_conflict_strategy, blocked_policy_or_approval,
blocked_external_ci, blocked_pr_metadata, blocked_artifact_context,
needs_user_decision, cleanup_only), and heartbeat, dirty-worktree,
validation, receipt, merge, cleanup, and blocker ledgers. Group repeats in
recurring_finding_classes with
finding_class_id, fingerprint_sha256, normalized_invariant, occurrences,
root cause, guardrail, and merge eligibility.
What ships with it
25 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.
- references/closeout-commander.md 10 KB
- references/contract.yaml 15 KB
- references/eval-scenarios.json 54 KB
- references/evals.yaml 52 KB
- references/evals/eval.harness.local-pass-ci-unknown.md 1.5 KB
- references/evals/eval.harness.pr-lifecycle-stops-before-main.md 1.7 KB
- references/evals/eval.pr-green-sweep.admin-merge-pressure.md 702 B
- references/evals/eval.pr-green-sweep.blocked-external-ci-keeps-independent-lanes-visible.md 1.2 KB
- references/evals/eval.pr-green-sweep.cleanup-without-merge-proof.md 1.6 KB
- references/evals/eval.pr-green-sweep.rotation-without-heartbeat.md 1.7 KB
- references/evals/eval.pr-green-sweep.stale-green-checks.md 1.6 KB
- references/evals/eval.pr-green-sweep.status-summary-without-queue.md 1.6 KB
- references/evals/eval.pr-green-sweep.wrong-validator-for-reference-doc.md 1.8 KB
- references/knowledge-capsule-routing.md 1.6 KB
- references/knowledge-capsule.manifest.yaml 9.0 KB
- references/knowledge-capsules/pr-green-sweep-action-queue.md 7.8 KB
- references/knowledge-capsules/pr-green-sweep-authorization-and-blockers.md 6.4 KB
- references/knowledge-capsules/pr-green-sweep-cleanup-proof.md 8.4 KB
- references/knowledge-capsules/pr-green-sweep-heartbeat-and-scope.md 8.8 KB
- references/knowledge-capsules/pr-green-sweep-live-pr-evidence.md 12 KB
- references/knowledge-capsules/pr-green-sweep-validation-surface.md 9.7 KB
- references/knowledge-demand.yaml 982 B
- references/recurring-finding-ledger.v1.schema.json 3.0 KB
- references/task-profile.json 1.4 KB
- scripts/validate_recurring_findings.py 2.8 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.
- 8d ago First seen · 204 lines · 51 tokens per session scan A 2eb34dbb30b7
pr-green-sweep is a skill published in the GitHub repository jscraik/Agent-Skills (8 stars, last pushed 10d ago), licensed Apache-2.0. It adds 51 tokens to every session and 2,532 once invoked, about $0.0003 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 2026-09-03.
Other skills, from other repositories
pr-review
Address feedback left on a GitHub pull request: fetch unresolved review threads, make agreed Elixir/Phoenix code fixes, reply, and resolve. Use for a PR URL/number or reviewer comments. NOT for pre-PR review, findings triage, or CI monitoring.
phx-pr-review
Address feedback left on a GitHub pull request: fetch unresolved review threads, make agreed Elixir/Phoenix code fixes, reply, and resolve. Use for a PR URL/number or reviewer comments. NOT for pre-PR review, findings triage, or CI monitoring.
merge-queue
Process the Refinery merge queue - collect agent work, detect and resolve conflicts, merge in dependency order, and verify integration.
pr-threshold
Track commit accumulation and trigger PR when thresholds crossed.
commit-message
Reviews working-tree changes, then drafts a Conventional Commits title/body and states the semantic-release version bump a single such commit would imply. Also notes which defensive-code categories were touched. Use when the user wants to commit recent work, prepare a Conventional Commits message, or asks for…
github-commenting
How to post clean, rich, deduplicated GitHub PR review comments — suggestion blocks, multi-line anchors, markers, formatting rules. Load before posting or fixing any PR comment.