seo-unlighthouse

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill delegates execution to scripts/unlighthouse_run.py, which manages the Unlighthouse CLI as a subprocess for auditing sites.
  • [EXTERNAL_DOWNLOADS]: Requires the installation of the unlighthouse-cli Node.js package and references an environment setup script (extensions/unlighthouse/install.sh).
  • [PROMPT_INJECTION]: Potential surface for indirect prompt injection from data audited on external websites.
  • Ingestion points: Data from external URLs is crawled and aggregated into the ci-result.json report file.
  • Boundary markers: No explicit markers or delimiters are mentioned for the report parsing phase.
  • Capability inventory: The skill uses subprocess execution (unlighthouse_run.py) and has file read/write access for audit reports.
  • Sanitization: Includes a url_safety pre-flight check to validate target URLs before the audit begins.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 02:37 AM
Security Audit — agent-trust-hub — seo-unlighthouse