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.
npx skills add Kerliula/ship-me --skill critique-megit clone --depth 1 https://github.com/Kerliula/ship-meWrote 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/kerliula/ship-me/critique-me)<a href="https://agentmods.dev/skills/kerliula/ship-me/critique-me"><img src="https://agentmods.dev/badge/skills/kerliula/ship-me/critique-me/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/kerliula/ship-me/critique-me"><img src="https://agentmods.dev/badge/skills/kerliula/ship-me/critique-me.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 | $0.00107 | $0.02307 |
| Opus 5.5 | $0.00043 | $0.00923 |
| Sonnet 5.5 | $0.00021 | $0.00461 |
| Haiku 4.5 | $0.00011 | $0.00231 |
Grade A, and why
critique-me 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 28d 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 — 244 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Critique Me — Try to Break the Solution
You are the reviewer, not a second designer. Your job is to find where
the solution is wrong, weak, or missing something — before anyone
approves it and before a line of code exists. A finding caught here
costs a paragraph. The same finding caught in /verify-me costs a
rebuild.
Assume the solution has at least one real problem and look for it. But never invent one. If a sub-problem holds up, say so in one line and move on. A critique padded with weak findings is as useless as a solution padded with fake options.
Input
You need two files:
- The solution:
docs/solutions/<slug>.md - The problem it answers:
docs/grilling/<slug>.md(its path is under "Source problem" in the solution file) - What can break it:
docs/breakers/<slug>.md
If either is missing, stop and say so. Don't review a solution against a problem you had to guess.
Read both fully before doing anything else.
Step 1 — Check it against the developer's answers
The grilling file is what the developer actually said. The solution has to honor all of it. Check:
- Every requirement is served. Each R-number is covered by at
least one sub-problem's
Serves:line, and the recommended option really does satisfy it (not just claims to). - Out of scope stays out. No recommended option builds something the developer ruled out.
- Every edge case is handled. Walk each one through the recommended options yourself. Don't trust the solution's own "Edge cases, checked" section — re-check it.
- Every "must always stay true" rule holds. Same: re-check it, don't copy the solution's verdict.
- Each recommendation is justified by the developer's facts. A
Why:line that rests on taste ("simpler", "cleaner") instead of a rule, limit, or edge case from the grilling file is a finding. - Load-bearing assumptions are respected. The solution must fit every 🔴 in the grilling file's "Load-bearing assumptions" table. If a recommended option would be hard to undo should an Accepted risk assumption turn out wrong, say so — and name a close option that would survive either answer, if one exists.
- Each
Rejected:line is fair. If an option lost for a reason that's wrong, or that applies just as much to the winner, say so.
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.
- 28d ago First seen · 244 lines · 107 tokens per session scan A 84064acfd422
critique-me is a skill published in the GitHub repository Kerliula/ship-me (1 stars, last pushed 19d ago), licensed MIT. It adds 107 tokens to every session and 2,307 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-19.
Other skills, from other repositories
laravel-architecture-review
Review a Laravel codebase's architecture and design, not just its types. Use when asked to audit a Laravel app, find architectural smells, decide where business logic belongs (controller vs Action vs Service), whether a Repository/DDD/TDD layer is justified, or how to apply SOLID/DRY to a Laravel project. Runs the…
loop-council
Sends an open design question to several independent voices in parallel and returns a decision with its evidence. Use before building anything that touches money, permissions, auth, or schema, or whenever the honest answer is "I am not sure".
react-patterns
React 19 performance patterns and composition architecture for Vite + Cloudflare projects. 50+ rules ranked by impact — eliminating waterfalls, bundle optimisation, re-render prevention, composition over boolean props, server/client boundaries, and React 19 APIs. Use when writing, reviewing, or refactoring React…
critical-code-reviewer
Rigorously review code or pull requests for correctness, security, accessibility, maintainability, tests, and edge cases. Use when users request a critical code review, want a guided walkthrough of findings, need implementer-facing feedback, or want to prepare, create, or submit a GitHub pull request review.
review-work
Post-implementation gate review: run manual QA on the real surface yourself, then launch ONE gate reviewer (never a panel) to audit goal, constraints, code quality, security, missed context, and QA evidence. Use before a PR handoff or when the user explicitly asks to review completed work.
one-way-door
Flags irreversible decisions before commit. Use for data models, infra, auth boundaries, API contracts, event schemas, CI/CD.