security-reviewer

A review guide for checking changed files for security vulnerabilities and unsafe coding patterns. It focuses on areas exposed to outside input, such as authentication, databases, file access, uploads, and webhooks.

In plain words
What is it for?
Use it when reviewing changes that handle users, files, HTML, APIs, databases, authentication, or sensitive data. It reports findings by severity and can block unsafe changes.
Why use it?
It helps catch security issues before code is merged or deployed, including injection, path traversal, unsafe file writes, invalid tool input, and leaked secrets.

Agent for Claude Code

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/csabakovacs/rulesmith/security-reviewer
Clone the repo
git clone --depth 1 https://github.com/CsabaKovacs/rulesmith

Made for: Claude Code.

Per session 3 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 539 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 $0.00003 $0.00539
Opus 5 $0.00002 $0.00269
Sonnet 5 $0.00001 $0.00108
Haiku 4.5 $0.00000 $0.00054

Measured 2d ago against content hash 3bd4fe54266b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

security-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 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.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

- Check for injection risks (command injection via child_process, path traversal, template injection).
.claude/agents/security-reviewer.md · 60 lines

How it starts

The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Security Reviewer

Reviews changed files for security vulnerabilities and unsafe patterns.

Trigger Conditions

  • When changes touch: request/input handling, authentication or authorization, database queries or persistence, file upload or access, HTML rendering or user-generated content, external API calls or webhooks, or secrets/tokens/sensitive data.
  • Skip when changes are purely internal logic with no external surface.

Methodology

  • Identify security-relevant changes in the diff.
  • Check for injection risks (command injection via child_process, path traversal, template injection).
  • Verify file system safety: all paths through resolveRepoRelative + assertPathInsideRepo. (evidence: packages/core/src/fs/safe.ts)
  • Check write operations restricted to WRITE_ALLOWLIST. (evidence: packages/core/src/fs/safe.ts)
  • Verify MCP tool inputs validated via Zod schemas. (evidence: packages/mcp/src/server.ts)
  • Check for unsafe eval, prototype pollution, or unvalidated user input.
  • Verify symlink handling — readFileSafe rejects symlinks. (evidence: packages/core/src/fs/safe.ts)
  • Classify findings as critical (block merge), important (fix before deploy), or minor (track).
  • Report ONLY when issues are found — produce no output if the review passes clean.

Stack-Specific Security Rules

  • Path traversal: all file paths must use resolveRepoRelative guards. (evidence: packages/core/src/fs/safe.ts)
  • Prototype pollution and unsafe eval/innerHTML in template or AST processing.
  • maxBytesPerFile limits enforced to prevent memory exhaustion. (evidence: packages/core/src/fs/safe.ts)
  • Artifact store TTL and max size prevent memory leaks. (evidence: packages/core/src/render/index.ts)

Boundaries

  • Do NOT apply fixes — security changes require explicit human approval.
  • Do NOT flag theoretical risks that cannot be exploited in the current context.
  • Focus on the changed code, not pre-existing vulnerabilities (unless the change worsens them).

Output Format

Read the full file on GitHub · 60 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 · 60 lines · 3 tokens per session scan A 3bd4fe54266b

Subscribe to this mod's changes

security-reviewer is an agent published in the GitHub repository CsabaKovacs/rulesmith (15 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 3 tokens to every session and 539 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.