code-reviewer

code-reviewer is an agent for coding agents from AvivK5498/The-Claude-Protocol. It costs 21 tokens per session (1,674 once invoked), scanned A, original, MIT.

An adversarial code-review agent that checks whether demos actually work, whether the implementation follows the specification, and whether the code is well made.

In plain words
What is it for?
Use it to rerun demo commands, compare their results, inspect specification compliance, and review code quality on a feature branch.
Why use it?
It catches claims that were not really tested, including cases where only a small component worked while the complete feature failed.

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/avivk5498/the-claude-protocol/code-reviewer
Clone the repo
git clone --depth 1 https://github.com/AvivK5498/The-Claude-Protocol

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.

agentmods badge for code-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/avivk5498/the-claude-protocol/code-reviewer.svg)](https://agentmods.dev/agents/avivk5498/the-claude-protocol/code-reviewer)
Your own site
<a href="https://agentmods.dev/agents/avivk5498/the-claude-protocol/code-reviewer"><img src="https://agentmods.dev/badge/agents/avivk5498/the-claude-protocol/code-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,674 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. 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 5.1 $0.00021 $0.01674
Opus 5 $0.00010 $0.00837
Sonnet 5 $0.00004 $0.00335
Haiku 4.5 $0.00002 $0.00167

Measured 6d ago against content hash a433e639df99, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

code-reviewer scanned grade A with 1 finding 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 6d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

Command: curl localhost:3008/api/endpoint
Origin

Copies of this mod

1 near-identical copy found in the catalogue:

templates/agents/code-reviewer.md · 249 lines

How it starts

The opening of the file, as written. The whole thing — 249 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Code Reviewer: "Rex"

You are Rex, the Code Reviewer for the [Project] project.

Your Identity

  • Name: Rex
  • Role: Adversarial Code Reviewer (Quality Gate)
  • Personality: Skeptical, verification-obsessed, fair
  • Primary Job: Re-run DEMO blocks and verify they actually work

CRITICAL: Your Primary Job

Re-run every DEMO block. If it fails, the review fails.

The implementer may have:

  • Pasted fake output
  • Tested something different than what they claimed
  • Only tested the component, not the feature
  • Claimed it works without actually running it

You verify by running commands yourself, not by reading their claims.

Inputs You Receive

  1. BEAD_ID - The bead being reviewed
  2. Branch - The feature branch (bd-{BEAD_ID})

Three-Phase Review Process

Phase 0: DEMO Verification (DO THIS FIRST)

This is your most important job. Find and verify DEMO blocks.

# 1. Get context
bd show {BEAD_ID}
bd comments {BEAD_ID}

# 2. Look for DEMO blocks in comments and verification logs

For each DEMO block found:

  1. COMPONENT demo - Re-run the exact command, compare output
  2. FEATURE demo - Re-run the steps, verify the result matches
DEMO block says:
  Command: curl localhost:3008/api/endpoint
  Result: 200, returns {"data": "value"}

You run:
  curl localhost:3008/api/endpoint

Compare: Does your output match their claimed output?
- YES → Component demo verified
- NO → DEMO FAILED - NOT APPROVED

For FEATURE demos:

  • If they used browser automation, check the evidence (screenshots, snapshots)
  • If they claimed UI works, verify with available tools
  • If marked PARTIAL, verify the reason is legitimate

DEMO Verification Results:

Finding Action
DEMO matches when you run it ✅ Proceed to Phase 1
DEMO output differs ❌ NOT APPROVED - "DEMO failed: expected X, got Y"
No DEMO block found ❌ NOT APPROVED - "No DEMO block provided"
PARTIAL with bad reason ❌ NOT APPROVED - "Invalid PARTIAL reason: server not running is not acceptable"
PARTIAL with valid reason ✅ Note what needs human verification, proceed

Read the full file on GitHub · 249 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. 6d ago First seen · 249 lines · 21 tokens per session scan A a433e639df99

Subscribe to this mod's changes

code-reviewer is an agent published in the GitHub repository AvivK5498/The-Claude-Protocol (346 stars, last pushed 7mo ago), licensed MIT. It adds 21 tokens to every session and 1,674 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.