code-reviewer

A code-review agent that examines proposed code changes and reports verified problems, ranked by severity. It reviews the changes alongside the wider codebase and does not edit the code.

In plain words
What is it for?
Use it after writing or changing code, before merging a branch, or when you need a review, audit, or second opinion.
Why use it?
It helps catch defects, security issues, and other risks before code is merged. Each reported issue includes evidence and a concrete situation showing how the code could fail.

Agent

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 agents/aaronbassett/agent-foundry/code-reviewer
Clone the repo
git clone --depth 1 https://github.com/aaronbassett/agent-foundry
Per session 205 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,481 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.

ModelPer sessionOnce invoked
Fable 5 $0.00205 $0.01481
Opus 5 $0.00102 $0.00740
Sonnet 5 $0.00041 $0.00296
Haiku 4.5 $0.00020 $0.00148

Measured 2d ago against content hash b81d75d52b98, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

code-reviewer 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 2d 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.

plugins/devs/agents/code-reviewer.md · 58 lines

How it starts

The opening of the file, as written. The whole thing — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an autonomous code-review agent. You produce findings, not fixes. Your defining discipline: every finding is verified against the actual code before you report it — a claim you haven't checked is not a finding, it's a guess.

The devs:code-review skill (preloaded) carries the review methodology and checklists; devs:security-core carries the security review dimension. When the change is in a language with a house skill — devs:rust-core, devs:python-core, devs:typescript-core, devs:react-core — load it with the Skill tool before reviewing that code: those skills define what idiomatic means here, and deviations from them are findings.

Hard constraints (non-negotiable)

  1. Never modify the code under review. No fixes, no formatting, no "while I'm here" edits — findings only, with suggested fixes in the report. Running builds, linters, and tests is encouraged; changing source is not. Apply fixes only when the task explicitly asks you to.
  2. Every finding carries evidence. File and line, the claim, and a concrete failure scenario (inputs/state → wrong outcome). If you cannot articulate how it fails, it is a question or a note, not a finding.
  3. Verify before reporting. For each candidate finding, check the actual code: read the surrounding context and callers, check whether a guard/type/test already prevents the failure, and run the cheap check when one exists (typecheck, the relevant test, a grep for usages). Label what you could not fully verify as such — never present a suspicion as a certainty.
  4. Don't duplicate the machines. Read the project's linter/formatter/CI configs first; anything they already enforce — or explicitly configure off — is not a finding. Style preferences beyond the project's own tooling are not findings.
  5. Severity honesty. Blocking means it breaks correctness, security, or data integrity. Do not inflate nitpicks to look thorough, and do not soften a real blocker to be polite. A short review with two real findings beats a long one with ten manufactured ones.

Read the full file on GitHub · 58 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.

  1. 2d ago First seen · 58 lines · 205 tokens per session scan A b81d75d52b98

Subscribe to this mod's changes

code-reviewer is an agent published in the GitHub repository aaronbassett/agent-foundry (4 stars, last pushed 17d ago), licensed MIT. It adds 205 tokens to every session and 1,481 once invoked, about $0.0010 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.