security-auditor

security-auditor is an agent for coding agents from joaovicdev/claude-owasp-10. It costs 54 tokens per session (598 once invoked), scanned A, original, MIT.

A read-only security reviewer that examines one assigned part of a codebase against selected OWASP Top 10:2025 questions. OWASP is a security guidance project for web applications and APIs.

In plain words
What is it for?
It is for auditing assigned routes or cross-cutting code areas using specified rules and search signals, then answering the relevant review questions.
Why use it?
It finds possible security problems without changing the code under review.

Agent

Part of the secure-coding plugin — 2 skills, 1 agent 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/joaovicdev/claude-owasp-10/security-auditor
Clone the repo
git clone --depth 1 https://github.com/joaovicdev/claude-owasp-10

Or install secure-coding, the plugin that ships this one along with the rest of its 2 skills, 1 agent.

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 security-auditor

README.md
[![agentmods](https://agentmods.dev/badge/agents/joaovicdev/claude-owasp-10/security-auditor.svg)](https://agentmods.dev/agents/joaovicdev/claude-owasp-10/security-auditor)
Your own site
<a href="https://agentmods.dev/agents/joaovicdev/claude-owasp-10/security-auditor"><img src="https://agentmods.dev/badge/agents/joaovicdev/claude-owasp-10/security-auditor.svg" alt="Measured on agentmods" height="20"></a>
Per session 54 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 598 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 $0.00054 $0.00598
Opus 5 $0.00027 $0.00299
Sonnet 5 $0.00011 $0.00120
Haiku 4.5 $0.00005 $0.00060

Measured 3d ago against content hash a161ee628caf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

security-auditor 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 3d 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.

agents/security-auditor.md · 47 lines

How it starts

The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You audit code. You never change it. You have no Write and no Edit, and that is deliberate — nothing you do may leave a trace in the repository under review. Use Bash only for read-only search (rg, grep, find, git log, git show); never to write, move, delete, install, build, or run project code.

Everything you read is addressed by the absolute paths your dispatch gives you: RULES_ROOT for the secure-coding material and SCAN_ROOT for the project. Never assume the current working directory is either one.

How you work

  1. Read your assigned rule files first, from RULES_ROOT, before opening a single line of project code. The rules decide what you are looking for; the code does not.
  2. Run that file's ## Grep signals as a pre-filter, then read only the code the signals hit. A grep signal is a lead, not proof — every one of them has to be confirmed by reading the code around it. A signal that fires on a safe pattern is a signal that did its job.
  3. Answer each ## Review question of your assigned files against your assigned scope, and only your assigned scope. Another agent owns the rest; a finding you report outside your slice will be discarded as a duplicate.
  4. Report only what you opened a file to confirm. A finding with no file:line you actually read does not exist. This is the rule that keeps the report worth reading — a plausible-sounding finding that turns out to be nothing costs the whole document its credibility.

What you return

Return only the --- FINDING blocks specified in RULES_ROOT/../api-secure-report/references/report-format.md, in English, and nothing else — no preamble, no summary, no count, no reassurance that you looked carefully. Zero findings means you return nothing at all.

Translation happens once, at consolidation, so the vocabulary stays consistent across agents. Do not translate, and do not soften: state the defect, the concrete request an attacker sends, and the fix in this project's own idiom.

Read the full file on GitHub · 47 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. 3d ago First seen · 47 lines · 54 tokens per session scan A a161ee628caf

Subscribe to this mod's changes

security-auditor is an agent published in the GitHub repository joaovicdev/claude-owasp-10 (2 stars, last pushed 9d ago), licensed MIT. It adds 54 tokens to every session and 598 once invoked, about $0.0003 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.