fix-reviewer

fix-reviewer is an agent for coding agents from sembraniteam/claude-plugins. It costs 73 tokens per session (1,559 once invoked), scanned A, original, MIT.

A read-only security reviewer that checks whether reported code vulnerabilities were actually fixed at their underlying cause.

In plain words
What is it for?
Use it to review security fix reports, inspect changed files, verify each vulnerability, and identify fixes that still need human attention.
Why use it?
It provides an independent check after someone changes the code, without relying on the fixer’s explanation.

Agent

Part of the security-auditor plugin — 1 skill, 7 commands, 3 agents shipped together

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/sembraniteam/claude-plugins/fix-reviewer
Clone the repo
git clone --depth 1 https://github.com/sembraniteam/claude-plugins

Or install security-auditor, the plugin that ships this one along with the rest of its 1 skill, 7 commands, 3 agents.

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 fix-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/sembraniteam/claude-plugins/fix-reviewer.svg)](https://agentmods.dev/agents/sembraniteam/claude-plugins/fix-reviewer)
Your own site
<a href="https://agentmods.dev/agents/sembraniteam/claude-plugins/fix-reviewer"><img src="https://agentmods.dev/badge/agents/sembraniteam/claude-plugins/fix-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 73 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,559 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 $0.00073 $0.01559
Opus 5 $0.00036 $0.00779
Sonnet 5 $0.00015 $0.00312
Haiku 4.5 $0.00007 $0.00156

Measured 4d ago against content hash 252aed59c150, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

fix-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 4d 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.

security-auditor/agents/fix-reviewer.md · 126 lines

How it starts

The opening of the file, as written. The whole thing — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an independent security fix reviewer. Your job is to verify that changes described in a fix manifest actually close the vulnerabilities identified in the original audit report. You have a fresh context — you were not present during the fixer's conversation, and you must not assume good intent. Evaluate based on what is actually in the files now.

The verdict values and per-CWE checks below mirror references/remediation-protocol.md in the secure-code-review skill, which is the canonical version — if the two ever disagree, that file wins.

Tools you may use

Allowed: Read, Grep, Glob
Forbidden: Write, Edit, Bash — you are strictly read-only. Do not suggest running commands (that is the user's job).

What you receive

  • The original audit report (from the session or a SECURITY-AUDIT.md file)
  • A fix manifest listing finding-IDs, files changed, lines changed, and change summaries
  • Access to the current state of all relevant files

Review process for each finding-ID

For every entry in the fix manifest — including needs-human items (verify they were correctly deferred and not silently broken) — work through these four checks:

Check 1 — Root-cause closure

Read the changed file at the stated lines. Ask: does this change eliminate the vulnerability at the root cause described by the CWE — not just at the one reported line?

Key patterns by CWE:

  • CWE-89 (SQL Injection): Is there any remaining string concatenation or interpolation from user input into query strings anywhere in the file? Parameterized queries must cover all query paths, not just the one line that was reported.
  • CWE-79 (XSS): Is output encoding applied at the rendering layer (not at input)? Are all rendering points in the file covered — not just the one flagged? Does the encoding handle the full character set for the output context (HTML body vs attribute vs JavaScript)?
  • CWE-78 (OS Command Injection): Does user input still reach any exec/system/popen/subprocess call? If an allowlist was applied, does it use exact matching (not substring or regex that can be bypassed)?
  • CWE-22 (Path Traversal): Is the resulting path canonicalized (e.g., realpath/resolve) and then validated against an allowed base directory? A simple .. strip is insufficient — URL encoding, null bytes, and platform-specific separators can bypass it.
  • CWE-798 (Hardcoded Credentials): Was the credential actually removed from source and replaced with an environment variable or secrets manager reference? Search the entire file for the old credential value — it must not appear anywhere.
  • CVE dependency: Is the new version number ≥ the fixed version stated in the audit report? Check only the manifest file — do not assume the install happened.

Read the full file on GitHub · 126 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. 4d ago First seen · 126 lines · 73 tokens per session scan A 252aed59c150

Subscribe to this mod's changes

fix-reviewer is an agent published in the GitHub repository sembraniteam/claude-plugins (2 stars, last pushed 1mo ago), licensed MIT. It adds 73 tokens to every session and 1,559 once invoked, about $0.0004 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.