Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add dgilford/ai-science-toolkit/plugin install ai-science-toolkitWrote 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/dgilford/ai-science-toolkit/ai-review)<a href="https://agentmods.dev/skills/dgilford/ai-science-toolkit/ai-review"><img src="https://agentmods.dev/badge/skills/dgilford/ai-science-toolkit/ai-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/skills/dgilford/ai-science-toolkit/ai-review"><img src="https://agentmods.dev/badge/skills/dgilford/ai-science-toolkit/ai-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 | $0.00098 | $0.01965 |
| Opus 5 | $0.00049 | $0.00983 |
| Sonnet 5 | $0.00020 | $0.00393 |
| Haiku 4.5 | $0.00010 | $0.00197 |
Grade A, and why
ai-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 11d 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 — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run me on the strongest model at high (or higher) reasoning effort. This skill spends its budget on breadth and depth of thinking, not on speed. If invoked on a weaker model, say so and recommend re-running with fable at high+ effort.
You are a senior software engineer conducting a full review: deep coding expertise, product-delivery judgment, and grounded, disciplined creativity. You find what others miss — but every finding and every idea is anchored to evidence in this repo. No generic advice. No hallucinated defects.
Governing principle
Orchestrate, never duplicate. Each lane below is owned by a tool that already does it best. ai-review's unique value is (1) running them in parallel, (2) the three lanes no other tool covers, and (3) one deduplicated, ranked synthesis. Never re-implement a delegated lane's logic yourself.
Live state
git rev-parse --show-toplevel 2>/dev/null && git status --short 2>/dev/null | head -20 || echo "(not a git repo — review the given path as-is)"
Scope
- Default: the whole repo (git-tracked files).
[path]: restrict to a subtree.--since <ref>: review only files changed since<ref>(cheaper, PR-shaped).--fix: after reporting, apply only the HIGH-confidence deterministic repairs (via/unstale --auto). Everything else stays advisory. Without--fix, edit nothing.
State the resolved scope in the report header before doing anything else.
Lanes — dispatch as a parallel subagent fan-out
Spawn the delegated lanes as concurrent subagents in a single message (multiple Task calls). Each subagent runs the owning skill/command over the resolved scope and returns its structured findings. Do not re-derive their logic here.
| Lane | Owner (subagent invokes) | ai-review adds |
|---|---|---|
| Correctness | /code-review |
fold into synthesis |
| Security | /security-review |
fold into synthesis |
| Staleness | /unstale (report mode) |
dispatch only — zero duplicated staleness logic |
| Over-baking | /overbaked |
dispatch |
| Claim rigor | /reviewer-2 (factual/scientific/quantitative claims) |
dispatch when the repo makes such claims |
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.
- 11d ago First seen · 136 lines · 98 tokens per session scan A 8d6d14ba2116
ai-review is a skill published in the GitHub repository dgilford/ai-science-toolkit (62 stars, last pushed 22d ago), licensed MIT. It adds 98 tokens to every session and 1,965 once invoked, about $0.0005 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 2026-08-30.
Other skills, from other repositories
kiss-dry-yagni
Principes KISS, DRY, YAGNI. Use when reviewing code quality or refactoring.
rubber-ducky
Use when you've planned a non-trivial change and are about to implement it, finished a complex or multi-file piece of work, just wrote tests, or are stuck on repeated failures — and any time the user says "rubber duck this", "rubber ducky", "get a second opinion", "sanity-check my plan", "poke holes in this", "what am…
debugging
Systematic root-cause debugging: reproduce, investigate, hypothesize, fix with verification. Use when asked to "debug this", "fix this bug", "why is this failing", "troubleshoot", or mentions errors, stack traces, broken tests, flaky tests, regressions, or unexpected behavior.
ia-c-systems
C patterns for systems code, libraries, and native extensions: module layout, function decomposition, status-enum errors, memory safety, undefined behavior, and performance measurement. Use when writing, reviewing, refactoring, or debugging C, working with malloc lifetimes, buffer overflows, sanitizers, or Valgrind…
ia-verification-before-completion
Enforces fresh verification evidence before any completion claim. Use when about to claim "tests pass", "bug fixed", "done", "ready to merge", handing off work, or before editing when a request has ambiguous scope.
audit-and-fix
Audit an existing codebase, then autonomously implement the fixes it judges worth making against your next goal — composing deep-dive → triage → prompt-pack → build-loop to turn "here's what's wrong" into verified, receipted local commits + an honest ledger. ALWAYS invoke when the user says any of "audit this repo and…