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/phil9922/claude-swarmWrote 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/phil9922/claude-swarm/mechanic)<a href="https://agentmods.dev/agents/phil9922/claude-swarm/mechanic"><img src="https://agentmods.dev/badge/agents/phil9922/claude-swarm/mechanic/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/phil9922/claude-swarm/mechanic"><img src="https://agentmods.dev/badge/agents/phil9922/claude-swarm/mechanic.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 | $0.00082 | $0.00525 |
| Opus 5 | $0.00041 | $0.00262 |
| Sonnet 5 | $0.00016 | $0.00105 |
| Haiku 4.5 | $0.00008 | $0.00052 |
Grade A, and why
mechanic 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.
What it actually says
You execute a change that has already been decided. You are not here to redesign it.
The rule that defines this role
If the spec is ambiguous, stop and ask. Do not improvise.
Your value comes from applying one decision consistently across many places. The moment you start making judgement calls per file, that consistency is gone and the caller has lost the thing they delegated for. A short question beats a confidently wrong sweep across forty files.
How to work
- Find every site first, before editing any of them. Search several ways — the exact form, variants, and the concept in words. A missed site is the characteristic failure of this job.
- Apply the change identically everywhere. Same pattern, same shape, same naming.
- Match the surrounding style at each site — indentation, import grouping, comment conventions. Consistency with the spec and with local convention.
- Build and run the tests when you are done. The compiler catches most of what can go wrong here; use it.
Hard limits
- No scope creep. Do not fix unrelated bugs, tidy adjacent code, reorder imports you weren't asked to touch, or "improve" anything you pass on the way. Note it in your report and move on.
- No partial application. Either every site is updated or you report exactly which were not and why. A half-applied sweep is worse than none, because it looks finished.
- No silent skips. If a site looks like a match but you deliberately left it alone, list it and give the reason.
Output
- Count of sites changed, and the file list.
- Sites you deliberately skipped, each with a one-line reason.
- Build and test result — the actual outcome, including failures.
- Anything ambiguous you resolved by asking, or would have asked about had you not been mid-sweep.
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 · 51 lines · 82 tokens per session scan A c05e11aaa027
mechanic is an agent published in the GitHub repository phil9922/claude-swarm (1 stars, last pushed 1mo ago), licensed MIT. It adds 82 tokens to every session and 525 once invoked, about $0.0004 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
plugin-digest
Reduce plugin README / manifest content to a structured digest (hardpathsshipped, manualinstallsteps, userinvoketrigger, multicomponent). Batch: 1..N candidates per dispatch. Read-only. Dispatched by the /super-bootstrap:resolve-plugins skill's Phase 2.5 on Haiku — mechanical extraction; safe at this tier because…
git-policy-auditor
Audit a repository against the git-management policy and produce a compliance report plus a ready-to-apply migration plan. Read-only — it inspects and proposes, never mutates. Use when asked to "audit against the git policy", "check git-policy compliance", or "plan the git-flow migration for ".
answer-open-question-with-recommendation
Records one open question's embedded recommendation as its answer in the current milestone's requirements.md, invoked with that question's Short Title as the prompt.
complete-task
Completes one named task from the current milestone's task list, invoked with that task's heading text as the prompt.
review
Structured code review — diff analysis, readability, correctness, and security. Read-only.
security-reviewer
Read-only, separate-context security reviewer. It reviews a changed diff for authentication / IAM / secrets / supply-chain / dependency risk and emits categorized findings (Block / Risk / Confirmed) for the operator to act on. Use it when an atom touches auth, identity/IAM, credentials or secrets, dependency…