skill-writing

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes standard local file operations such as mkdir, ln, chmod, and test to establish a project-local skill structure and create discovery symlinks within .claude/skills/. These actions are scoped to the invocation directory and follow standard development workflows.\n- [EXTERNAL_DOWNLOADS]: Fetches documentation and formatting specifications from agentskills.io and code.claude.com (Anthropic). These downloads are used as reference material to ensure compliant skill authoring and do not include the execution of untrusted remote content.\n- [REMOTE_CODE_EXECUTION]: Recommends a deterministic execution model for scaffolded skills by using bun run for TypeScript and uv run for Python. This establishes a boundary between the model's stochastic prompt instructions and reproducible code logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 05:15 PM
Security Audit — agent-trust-hub — skill-writing