Borrowing it
Nothing to install: this file belongs to statecrafting/spec-spine. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/statecrafting/spec-spine/main/.claude/skills/shepherd/SKILL.mdgit clone --depth 1 https://github.com/statecrafting/spec-spineWrote 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/statecrafting/spec-spine/shepherd)<a href="https://agentmods.dev/skills/statecrafting/spec-spine/shepherd"><img src="https://agentmods.dev/badge/skills/statecrafting/spec-spine/shepherd/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/statecrafting/spec-spine/shepherd"><img src="https://agentmods.dev/badge/skills/statecrafting/spec-spine/shepherd.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 | $0.00060 | $0.01597 |
| Opus 5 | $0.00030 | $0.00798 |
| Sonnet 5 | $0.00012 | $0.00319 |
| Haiku 4.5 | $0.00006 | $0.00160 |
Grade A, and why
shepherd 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 yesterday.
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 — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/shepherd: from open PR to merged, with evidence
CI is where reality pushes back. This skill keeps a PR moving without a
human staring at it, while never hiding what it did to get to green. It
mirrors an orchestrator's shepherd stage: watch by head sha, remediate
through the gate, bounded attempts, honest stops. Bound by
.claude/rules/orchestrator-rules.md (checkpoints are real stops) and
.claude/rules/adversarial-prompt-refusal.md (a coupling refusal goes to
a human, not around them).
Step 0: resolve the PR
git branch --show-current
gh pr view ${ARGUMENTS:-} --json number,url,headRefName,headRefOid,baseRefName,mergeStateStatus,isDraft,reviewDecision
- No PR for the branch: stop and say so (run
/shipfirst). Do not open one here. - The branch must be a feature branch, never the default branch.
- Record
headRefOid: every check you read is for this sha and no other.
Step 1: watch the checks (by head sha)
Poll loop discipline: start at 15 s, multiply by 1.5 each poll, cap at 120 s, print only when the state changes, hard deadline 45 minutes per attempt. Never busy-loop.
gh pr checks <number> --json name,state,bucket,link 2>/dev/null \
|| gh pr view <number> --json statusCheckRollup
Read the rollup for the recorded head sha only. Classify:
- every required check
SUCCESS: go to Step 3; - any required check
FAILUREorCANCELLED: go to Step 2; PENDINGorQUEUED: keep polling;- a response missing the fields you need: stop as a typed failure ("cannot read a shape that cannot terminate"), never poll blind;
- the deadline passes: stop, report which checks never completed, and say the PR needs a human.
Step 2: remediate (at most two rounds)
For the failing check, fetch the evidence first:
gh run list --branch "$(git branch --show-current)" --limit 5 --json databaseId,name,conclusion,headSha
gh run view <run-id> --log-failed | tail -80
Then, on the branch:
- Diagnose from the log tail; read the failing test or gate output, not the summary line.
- Fix inside the spec's territory. If the failure is the coupling gate
(
C-001orC-002), claim the path in the spec being implemented or add theextendsedge on its owner; never edit an owning spec to ratify code that contradicts it (coherence guard: stop and report). A// Spec:comment header is the other legitimate claim for a file no frontmatter should list. - Run the gate exactly as
AGENTS.mdlists it; it must exit 0 locally. /commit(conventional, spec id as scope, regenerated shards staged), thengit push.- Re-read
headRefOid. Restart Step 1's watch on the new sha with a fresh 45 minute budget. Checks from the old sha are stale; never mix them.
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.
- yesterday First seen · 154 lines · 60 tokens per session scan A 3b0490684e0e
shepherd is a skill published in the GitHub repository statecrafting/spec-spine (10 stars, last pushed today), licensed Apache-2.0. It adds 60 tokens to every session and 1,597 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 2026-09-07.
Other skills, from other repositories
github
GitHub via gh CLI: PRs, issues, reviews, repos, auth.
github
Full GitHub CLI control — issues, PRs, code reviews, repo management. Uses gh CLI with auth detection, rate limiting, and templates. Triggers on: github, issue, pull request, PR, code review, repo, branch, label, assignee, milestone, release, workflow, actions.
pre-publish-review
Nuclear-grade 12-agent pre-publish release gate. Runs /get-unpublished-changes to detect all changes since last npm release, spawns up to 10 ultrabrain agents for deep per-change analysis, invokes /review-work (orchestrator manual QA plus one gate reviewer) for holistic review, and 1 oracle for overall release…
parallel-pr-review
Use when asked to "review the open PRs", review a batch or stack of pull requests, or run a recurring PR-review pass on a repo — especially with many PRs, stacked branches, conflicts, or security-sensitive changes. Covers grouping, fan-out to review subagents, verdict synthesis, and posting.
release
Release gentle-pi through GitHub and npm. Trigger: release, publish, npm publish, GitHub release, version bump.
oss-maintainer
Run an open-source project's issue/PR/release loop like a careful human maintainer — triage to root cause, absorb community PRs before duplicating them, gate every merge, ship honest releases, and thank the people doing your QA for free.