portfolio-taxonomy

Pass

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple bash scripts located in the plugin's local directory (e.g., clone-taxonomy.sh, validate-taxonomy.sh, edit-taxonomy.sh). It uses a dynamic filesystem path to locate these scripts within the user's home directory. This is standard behavior for the plugin's taxonomy management functions.
  • [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by allowing the import of external taxonomy definitions. Evidence: (1) Ingestion points: references/import-mode.md (Step 1) reads local JSON, CSV, or Markdown files. (2) Boundary markers: Absent. (3) Capability inventory: Executes bash scripts, performs file writes, and interacts with the user. (4) Sanitization: Post-import structural validation is conducted via validate-taxonomy.sh to ensure data integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 25, 2026, 01:51 PM
Security Audit — agent-trust-hub — portfolio-taxonomy