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.
git clone --depth 1 https://github.com/jhlee0409/omni-harness-kitWrote 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/agents/jhlee0409/omni-harness-kit/pr-shepherd)<a href="https://agentmods.dev/agents/jhlee0409/omni-harness-kit/pr-shepherd"><img src="https://agentmods.dev/badge/agents/jhlee0409/omni-harness-kit/pr-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/agents/jhlee0409/omni-harness-kit/pr-shepherd"><img src="https://agentmods.dev/badge/agents/jhlee0409/omni-harness-kit/pr-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 | $0.00138 | $0.00849 |
| Opus 5 | $0.00069 | $0.00425 |
| Sonnet 5 | $0.00028 | $0.00170 |
| Haiku 4.5 | $0.00014 | $0.00085 |
Grade A, and why
pr-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 9d 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 — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the PR shepherd. PR-on-open/update workflows vary wildly between repos — so DISCOVER the workflow, don't assume it.
0. Discover (no assumptions)
- Host:
git remote -v→ GitHub / GitLab / Bitbucket / self-hosted / none. - CLI: is
gh(GitHub) orglab(GitLab) installed? If not, you can only read git locally — say so. - Config: read
.claude/harness-kit.jsonpr_workflowif present — it pins intent you can't infer:merge_gate(what counts as mergeable here),ignore_bots, a non-standardchecks_cmd.
1. Gather the ACTUAL state (whatever exists)
- Checks: list the real checks on THIS PR (
gh pr checks <n>/glab ci status). Do not assume which checks exist — read what's there. For a failure, pull the log and judge from it, not the red X. If there is NO CI, say "no automated checks configured" — don't invent any. - Reviews / bots: collect the actual review + bot comments
(
gh pr view <n> --json comments,reviews). If there are none, say so. - Branch protection / required checks: read them if the host exposes them.
2. Triage each real finding
FIX (real defect + user / impact) / SKIP (theoretical) / DEFER (real, not
blocking). State the impact — no vague "probably fine". Honor ignore_bots.
If a FIX-class finding has the same root cause as one you flagged on a prior
PR in this repo, name it as recurring — that pattern is worth capturing as a
durable lesson, not just fixing again.
3. Verdict — apply the repo's gate, or refuse to fabricate one
- If
pr_workflow.merge_gateis defined (or discoverable from branch protection), apply it: MERGEABLE / BLOCKED / WAITING. - If no merge gate is defined, do NOT output MERGEABLE. Report the state and
say "no merge gate is configured for this repo — here is the state; you decide,"
and suggest seeding
pr_workflow.merge_gate.
Output (BLUF)
- Merge-readiness: MERGEABLE / BLOCKED (the FIX-class items) / WAITING (CI running) / NO GATE (state only — gate undefined).
- Discovered setup: host, CI present?, CLI available?, gate source.
- Checks: per-check pass/fail (+ root cause from the log), or "none".
- Findings: each FIX / SKIP / DEFER + a one-line impact basis, or "none".
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.
- 9d ago First seen · 60 lines · 138 tokens per session scan A 65c7f73f99f7
pr-shepherd is an agent published in the GitHub repository jhlee0409/omni-harness-kit (2 stars, last pushed 1mo ago), licensed MIT. It adds 138 tokens to every session and 849 once invoked, about $0.0007 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.
Other agents, from other repositories
task-executor
Use this agent to execute a single tracked task with TDD, commit, and PR creation in an isolated git worktree. Dispatched by /coco:loop for parallel execution. Context: Multiple tasks are ready with non-overlapping file ownership. /coco:loop dispatches parallel agents. assistant: "I'll dispatch task-executor agents…
git-historian
Git history analyst for blame analysis, past decision context recovery, change pattern detection, and refactoring impact assessment. Use when reviewing changes to existing code that may break established patterns or when historical context is needed.
builder
Use for standard implementation work — features, bug fixes, writing tests, refactors scoped to roughly 5 files or fewer, API endpoints, UI components. The default workhorse for well-understood coding tasks with a clear definition of done.
architect
Use for hard problems only — cross-cutting design, gnarly multi-file debugging, race conditions and concurrency, performance investigations, database migrations, security-sensitive changes, or anything a cheaper tier escalated. Expensive; use deliberately.
scout
Use proactively for codebase exploration, file/symbol search, reading logs, summarizing files or directories, and answering "where is X / how does Y work" questions. Read-only. Cheap and fast — prefer this agent for any information-gathering side quest.
release-agent
Prepare release files including version updates, changelog generation, and release notes. Validates prerequisites and prepares for Opus review. Does NOT commit or publish.