tk-setup-skills

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes metadata from external TigerKit skill files. It incorporates strong defensive measures including: 1) Ingestion points: Reads local files from paths provided to the planner script. 2) Boundary markers: Rejects path traversal attempts and all symlinks to ensure file access stays within defined roots. 3) Capability inventory: Prepares shell command strings but instructs the agent not to execute them. 4) Sanitization: Implements a custom YAML loader based on SafeLoader to mitigate parsing vulnerabilities.
  • [EXTERNAL_DOWNLOADS]: Mentions the use of 'npx skills@1.7.4' for obtaining skill packages, which is standard behavior for its stated purpose.
  • [COMMAND_EXECUTION]: Uses a local Python script for validation tasks; the script is carefully written to be read-only and validates all inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 12:27 AM