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 arcasilesgroup/ai-engineering --skill review-routergit clone --depth 1 https://github.com/arcasilesgroup/ai-engineeringWrote 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/arcasilesgroup/ai-engineering/review-router)<a href="https://agentmods.dev/skills/arcasilesgroup/ai-engineering/review-router"><img src="https://agentmods.dev/badge/skills/arcasilesgroup/ai-engineering/review-router/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/arcasilesgroup/ai-engineering/review-router"><img src="https://agentmods.dev/badge/skills/arcasilesgroup/ai-engineering/review-router.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 | $0.00137 | $0.02282 |
| Opus 5.5 | $0.00055 | $0.00913 |
| Sonnet 5.5 | $0.00027 | $0.00456 |
| Haiku 4.5 | $0.00014 | $0.00228 |
Grade A, and why
review-router 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 1mo 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 — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Route the review
full-review spawns up to five reasoning agents that each read a real amount of code. On a
four-line diff that is more expensive than the change was, and the extra lanes return
nothing — which trains people to stop running reviews at all. This skill makes the lane
selection a decision based on evidence in the diff, rather than a habit.
You decide which lanes run and how deep. You perform no review yourself.
Two rules bound everything below:
- Never route to zero lanes on a diff that touches first-party source. If the change is
genuinely trivial, that is
build-checkplus a targetedcode-audit, not silence. - Name what you skipped and why, in the output. A lane that was never run must never
read as a lane that came back clean. This is the same failure
full-reviewguards against when an agent dies.
1. Resolve scope and stack once
Per .claude/review/CONVENTIONS.md §1 and §3. Resolve both here, exactly once, and pass
the identical strings to every agent you spawn. Agents that resolve their own scope drift
apart and their findings stop being comparable — and in a monorepo they can legitimately
land on different stacks, which silently produces reviews of different things.
2. Measure the diff before you classify it
Get the shape of the change from git, not from a guess:
git diff <base>...HEAD --stat
git diff <base>...HEAD --name-only
Record: files changed, lines added/removed, and which subsystems are touched. Then subtract
everything CONVENTIONS.md §3 excludes — lockfiles, generated clients, build output,
vendored code. A 4,000-line diff that is 3,900 lines of lockfile is a 100-line diff,
and misreading that is the most common way a router over-spends.
If, after exclusions, the change touches no first-party source at all — docs only, config
comments, a lockfile refresh — say so and stop. Offer build-check alone. Do not spawn
reasoning agents to review prose.
3. Map evidence to lanes
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.
- 1mo ago First seen · 159 lines · 137 tokens per session scan A d392bfc8fc8d
review-router is a skill published in the GitHub repository arcasilesgroup/ai-engineering (61 stars, last pushed 2d ago), licensed Apache-2.0. It adds 137 tokens to every session and 2,282 once invoked, about $0.0005 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 2026-09-12.
Other skills, from other repositories
code-review
Review pull requests for repository fit, meaningful value, trustworthy provenance, differentiation, and maintainability in awesome-copilot.
spec-kitty-mission-review
Review a fully merged Spec Kitty mission post-merge (all WPs done/approved) to verify spec→code fidelity, FR coverage, drift, risks, and security. Triggers: "review the merged mission", "post-merge mission review", "verify the completed mission", "audit the mission implementation", "mission-level acceptance review"…
code-review-csharp
Perform structured code reviews of C# source code covering naming conventions, performance, security, readability, and .NET best practices. Trigger phrases include "review this C# code", "check my C# for best practices", "analyze this C# class", "find issues in my C# code".
mcp-server-review
Review a Model Context Protocol (MCP) server implementation against the 2026-07-28 protocol revision, which removed the initialize handshake and Mcp-Session-Id and made the protocol stateless. Use when the user asks to review, audit, upgrade or migrate an MCP server, asks whether their MCP server is spec compliant…
spec-kitty-runtime-review
Review runtime-owned outputs using the Spec Kitty review workflow surface, then direct approval or rejection with structured feedback. Triggers: "review this work package", "check runtime output", "approve this step", "review WP", "is this WP ready to approve", "check this implementation". Does NOT handle: setup-only…
adversarial-squad
Deploy a bounded, profile-loaded adversarial review squad at an SDD point-cut (post-spec, post-plan, post-tasks, pre-merge, or an ad-hoc decision) so independent doctrine lenses converge on findings one reviewer would miss. Triggers: "deploy a squad", "adversarial squad", "post-tasks anti-laziness pass", "pre-spec…