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/karsten-s-nielsen/mad-scientist-skillsWrote 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/commands/karsten-s-nielsen/mad-scientist-skills/re-review)<a href="https://agentmods.dev/commands/karsten-s-nielsen/mad-scientist-skills/re-review"><img src="https://agentmods.dev/badge/commands/karsten-s-nielsen/mad-scientist-skills/re-review/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/commands/karsten-s-nielsen/mad-scientist-skills/re-review"><img src="https://agentmods.dev/badge/commands/karsten-s-nielsen/mad-scientist-skills/re-review.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 | $0.00020 | $0.00456 |
| Opus 5 | $0.00010 | $0.00228 |
| Sonnet 5 | $0.00004 | $0.00091 |
| Haiku 4.5 | $0.00002 | $0.00046 |
Grade A, and why
re-review 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 12d 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
Re-review a revised artifact.
Artifact: $ARGUMENTS
Use the unbiased-review skill. Read references/re-review.md first, then the reference for the
artifact's type and references/output-contract.md.
Two jobs, one prohibition. Confirm the revisions were actually made; catch anything the revisions broke; do not re-litigate what the last round settled.
Start by reading the prior report in reviews/ in full — every finding ID, its severity, its
Resolved when condition, the recorded HEAD and artifact sha256, and anything explicitly accepted
or declined. Then compute what changed before judging anything. If the artifact is unversioned, try to
recover the reviewed revision from loose git objects so the comparison is a real line-level diff rather
than a section-level guess; if it cannot be recovered, say so and treat it as a limitation of the round.
Verify each prior finding from the tree with fresh evidence. Never accept the author's claim that
they fixed it — independent confirmation is the entire point of the round. A fix that satisfies the
letter of Resolved when but not the defect is PARTIAL; say which. If the condition could be met
without fixing the problem, the condition was badly written, and that is worth reporting too, plainly
and as your own error.
New findings are admissible only if they are in changed content, or a prior judgement whose basis the change invalidated. Anything else goes in a one-line Outside this round note.
Carry the settled ledger forward — resolved findings, accepted decisions, and declined CONSIDERs are
closed. DECLINED is durable; re-raising it is a form of not listening.
End with the explicit regression line. If everything blocking is resolved and nothing new turned up,
say APPROVE plainly — a round that manufactures a finding to justify itself is worse than one that
says "ship it".
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.
- 12d ago First seen · 37 lines · 20 tokens per session scan A d75a50a2fb4f
re-review is a command published in the GitHub repository karsten-s-nielsen/mad-scientist-skills (3 stars, last pushed 14d ago), licensed MIT. It adds 20 tokens to every session and 456 once invoked, about $0.0001 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 commands, from other repositories
merge
Finish a PR properly: every check green, every review addressed — human and bot — then merge and clean up.
pr
Prepare and open a pull request the senior way: gate, template, scrubbed, everything visible.
spec
Spec-first design: a gap-closing interview that produces a complete spec, with a quality controller that blocks until every section is answered and every question resolved.
debug
Systematic debugging: root cause before any fix, one hypothesis at a time, and a three-strikes rule that questions the architecture instead of stacking patches.
plan
Turn an approved spec into an implementation plan an engineer with zero context could execute — with a quality controller that blocks placeholders and hollow tasks.
tdd
Test-driven development with tests that actually catch breaks: red before green, name the break each test catches, and the mutation check before done.