skill-scanner-test CLAUDE.md

skill-scanner-test CLAUDE.md is an instructions file for coding agents from AppSecHQ/skill-scanner-test. It costs 902 tokens per session, scanned A, original, MIT.

Project instructions for a security research tool that scans popular AI-agent skills for weaknesses such as prompt injection, data theft, and malicious code.

In plain words
What is it for?
Use them when running the skill-scanner project, including identifying, cloning, scanning, and documenting AI-agent skills.
Why use it?
They keep the investigation ordered and require checking the first skill completely before expanding the scan to others.

Instructions file

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 instructions/appsechq/skill-scanner-test/claude-md
Clone the repo
git clone --depth 1 https://github.com/AppSecHQ/skill-scanner-test

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 skill-scanner-test CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/appsechq/skill-scanner-test/claude-md.svg)](https://agentmods.dev/instructions/appsechq/skill-scanner-test/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/appsechq/skill-scanner-test/claude-md"><img src="https://agentmods.dev/badge/instructions/appsechq/skill-scanner-test/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 902 This file is loaded in full into every session.
When invoked 902 The same file — it is already loaded in full.
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 $0.00902 $0.00902
Opus 5 $0.00451 $0.00451
Sonnet 5 $0.00180 $0.00180
Haiku 4.5 $0.00090 $0.00090

Measured 4d ago against content hash 9798b9c408db, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-scanner-test CLAUDE.md 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.

CLAUDE.md · 99 lines

How it starts

The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md - Skill Scanner Project

This file provides guidance to Claude Code when working on this security research project.

Project Overview

This project scans the top N most-installed AI agent skills from skills.sh using Cisco's skill-scanner tool to identify security vulnerabilities, prompt injection risks, and malicious patterns.

Objective

Produce a comprehensive security assessment report of popular AI agent skills to identify:

  • Prompt injection vulnerabilities
  • Data exfiltration patterns
  • Malicious code patterns
  • Other security concerns

Approach

IMPORTANT: Iterative execution with user checkpoints

  1. Perform a complete end-to-end test on the #1 top skill first (all phases: identify, clone, scan, document)
  2. Get explicit confirmation from the user before proceeding to skill #2
  3. Only after user approval, continue with subsequent skills

This approach validates the entire workflow before scaling up.

Directory Structure

skill-scanner/
├── CLAUDE.md              # This file - project instructions
├── PROJECT-PLAN.md        # Detailed execution plan
├── design-plan/           # Internal design docs & plans (gitignored)
├── skills/                # Cloned skill repositories
│   └── skill-inventory.md # Inventory of top 25 skills with metadata
└── results/               # Scan output files
    ├── <skill>-scan.json  # Raw JSON scan results
    ├── <skill>-scan.md    # Human-readable scan results
    └── summary-report.md  # Consolidated findings report

Key Resources

Execution Phases

Phase 1: Environment Setup

  1. Install cisco-ai-skill-scanner via pip
  2. Verify installation with skill-scanner --help

Phase 2: Identify Top Skills

  1. Scrape or browse https://skills.sh/ to identify top skills by install count
  2. Document each skill in skills/skill-inventory.md with:
    • Skill name
    • GitHub repo URL
    • Install count
    • Brief description

Read the full file on GitHub · 99 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 · 99 lines · 902 tokens per session scan A 9798b9c408db

Subscribe to this mod's changes

skill-scanner-test CLAUDE.md is an instructions file published in the GitHub repository AppSecHQ/skill-scanner-test (5 stars, last pushed 3mo ago), licensed MIT. It adds 902 tokens to every session, about $0.0045 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.