A risk-based code-review command that routes review effort according to how much damage a changed area could cause. It follows the project's review rules instead of checking every file equally.

Use it to review current changes from a chosen Git base, identify failure modes and required guard tests, or audit unchanged critical infrastructure.

It focuses attention on critical defects in areas such as audit controls, authentication, human approval steps, and rate limits, which a broad scan may miss.

review — Command for Claude Code by heznpc: A risk-based code-review command that routes review effort according…

Install

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.

agentmods

npx agentmods add commands/heznpc/airmcp/review

Clone the repo

git clone --depth 1 https://github.com/heznpc/AirMCP

Made for: Claude Code.

Or install airmcp, the plugin that ships this one along with the rest of its 1 command, 1 MCP server.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/heznpc/airmcp/review.svg)](https://agentmods.dev/commands/heznpc/airmcp/review)

Your own site

<a href="https://agentmods.dev/commands/heznpc/airmcp/review"><img src="https://agentmods.dev/badge/commands/heznpc/airmcp/review.svg" alt="Measured on agentmods" height="20"></a>

Per session 23 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 585 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00023	$0.00585
Opus 5	$0.00012	$0.00293
Sonnet 5	$0.00005	$0.00117
Haiku 4.5	$0.00002	$0.00059

Measured 5d ago against content hash 8088b2b11bbd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

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 5d 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.

.claude/commands/review.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run a risk-tier-routed review of the current changes, per docs/rfc/0013-review-process.md. A flat scan over 272 tools dilutes attention and misses the silent-systemic defects (audit seal / HITL bypass / OAuth / rate-limit) that actually matter. Route effort to blast radius instead.

Step 1 — get the routing plan (don't skip; this is the mechanism)

Run the executable router and read its output:

node scripts/review-route.mjs --base "${ARGUMENTS:-origin/main}"

It prints, for the diff: each changed file's tier, the failure modes to hunt per touched T0/T1 area, the guard tests that must stay green, and a ⚠ warning for any T0 file changed without touching its guard test.

(For a periodic audit of unchanged infra — RFC 0013 §5 — run node scripts/review-route.mjs --audit instead and review every T0 area.)

Step 2 — review at the routed depth, NOT uniformly

T0 (critical infra) — review at MAX depth. For each touched area, hunt the specific failure mode the router named (e.g. audit → seal/_prev mismatch + tamper-not-detected; OAuth → alg=none/HS confusion + scope bypass). A generic "looks fine" is not a T0 review. Confirm the named guard tests cover the change; if a T0 file changed with no guard-test change, treat closing that gap as higher priority than the feature itself.
T1 (system / ui / finder / shortcuts) — the agent-drives-the-Mac surface. Verify HITL coverage, destructiveHint annotations, scope-gate mapping, and rate-limit tier (RFC 0014 §4.5).
T2 (JXA-thin modules) — light. Check esc()/escaping, the okUntrusted/toolError result shape, and that the contract test still passes. Don't spend max effort here; it fails loud, blast radius is one call.
T3 (generated / vendored) — review the generator (e.g. scripts/gen-swift-intents.mjs) and the drift guard, never the generated output.

Step 3 — report

Lead with the highest tier touched and the T0/T1 findings against the hunt list. Note any unguarded-T0 gap explicitly. Keep T2/T3 to a line unless the contract tests fail.

Read the full file on GitHub · 48 lines

Changes

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.

5d ago First seen · 48 lines · 23 tokens per session scan A 8088b2b11bbd

Subscribe to this mod's changes

review is a command published in the GitHub repository heznpc/AirMCP (16 stars, last pushed 2d ago), licensed MIT. It adds 23 tokens to every session and 585 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 2026-08-30.

Code review ai apple applescript automation calendar claude

review

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

Step 1 — get the routing plan (don't skip; this is the mechanism)

Step 2 — review at the routed depth, NOT uniformly

Step 3 — report

What this file has done since we first saw it

Other commands, from other repositories

pear-setup

briefing

schedule

merge-and-status

webhooks

templates