write-a-skill

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references documentation and templates from authoritative sources, including Anthropic's official agent documentation and public repositories for attribution and pattern reference.
  • [COMMAND_EXECUTION]: Provides local Python validation scripts intended to be run by the agent or user to verify skill integrity. These scripts (e.g., skill_description_validator.py, skill_structure_validator.py) rely solely on the Python standard library and perform local file analysis.
  • [INDIRECT_PROMPT_INJECTION]: The skill contains an attack surface for indirect injection as it processes user requirements and external documents to author new skills.
  • Ingestion points: User-provided task descriptions and document paths defined in SKILL.md.
  • Boundary markers: Not explicitly implemented in the authoring instructions.
  • Capability inventory: Local file system access for writing new skill files and execution of included validation scripts.
  • Sanitization: The skill does not define specific sanitization routines for the data it processes into generated skills.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:45 PM
Security Audit — agent-trust-hub — write-a-skill