skills.rest
stefaniuk
stefaniukCOMMUNITY
Explore all skills by this author →
GitHub

skill-scanner

Scan agent skills for prompt injection, malicious code, and excessive permissions.

1|Updated Jan 4, 2026
Search Tags:#static analysis#security#audit#permissions#prompt injection#supply chain#skill scanner
One-click install
npx skills add https://github.com/stefaniuk/loadout --skill skill-scanner-stefaniuk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-scanner
Source: https://github.com/stefaniuk/loadout/tree/main/.github/skills/skill-scanner
Command: npx skills add https://github.com/stefaniuk/loadout --skill skill-scanner-stefaniuk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Installing third-party agent skills without review exposes your environment to prompt injection, credential theft, data exfiltration, and hidden code execution. This Skill audits a skill directory before adoption so you can decide whether it is safe to install. ## Core Features & Use Cases - Static Security Scanning: Runs a bundled Python scanner that detects prompt injection patterns, obfuscation (zero-width characters, Unicode tag smuggling, base64), hardcoded secrets, dangerous code patterns, and structural attacks like symlinks, frontmatter hooks, npm lifecycle scripts, and PNG metadata injection. - Behavioral and Permission Analysis: Guides the agent through frontmatter validation, description-vs-instructions alignment checks, least-privilege tool assessment, and supply chain review of URLs and dependencies. - Structured Risk Report: Produces a severity-ranked findings report with confidence levels, evidence, remediation advice, and a final install recommendation. - Use Case: Before installing a community skill, ask the agent to scan its directory; the Skill runs the scanner, reviews scripts and references for malicious intent, and returns a risk rating from Clean to Critical. ## Quick Start Scan the skill in .agents/skills/my-new-skill for security issues and tell me whether it is safe to install.

Frequently Asked Questions about skill-scanner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scan an agent skill for security issues?▼

Point the Skill at the skill directory containing SKILL.md and ask for a security scan. It runs the bundled scan_skill.py static analyzer, then performs behavioral, permission, and supply chain analysis before returning a risk-rated report.

What does the skill security scanner detect?▼

It detects prompt injection patterns, obfuscation like zero-width characters and Unicode tag smuggling, hardcoded secrets, dangerous code such as eval and reverse shells, symlinks escaping the skill directory, frontmatter hooks, npm lifecycle scripts, and hidden text in PNG metadata.

Does the scanner require any tools to be installed?▼

Yes, the bundled Python script is executed with the uv CLI, which must be installed beforehand. The script itself declares pyyaml as its only dependency via PEP 723 inline metadata, which uv resolves automatically.

Why does the scanner flag security reference files as suspicious?▼

Security skills legitimately document attack patterns, so pattern matches in reference files are expected false positives. The workflow explicitly requires evaluating intent: patterns that document threats are legitimate, while patterns that execute against the agent are malicious.

Can the scanner catch attacks that are not in the text content?▼

Yes, it checks structural attack vectors including symlinks resolving outside the skill directory, YAML frontmatter hooks that auto-execute shell commands, test files auto-discovered by pytest or Jest, npm postinstall scripts, and text hidden in PNG metadata chunks.

Related Skills

Explore More Software Engineering Skills→
hepai-labhepai-lab
21

download_github_skills

Download a targeted skill folder from a GitHub repository using sparse-checkout.

Community
Intermediate
agulliagulli
482

skill-security-audit

Audit SKILL.md files for malicious, misconfigured, or untrusted content.

Community
Advanced
heitorfreitasferreiraheitorfreitasferreira
1

docker-security-guide

Apply Docker security guidelines for image, build, runtime, and network hardening.

Community
Advanced
OpenHandsOpenHands
134

docker

Start a Docker daemon and run containers inside containerized environments.

Official
Advanced