seo-unlighthouse
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
unlighthouseCLI tool and helper Python scripts (unlighthouse_run.py,lcp_subparts.py) through a wrapper to perform website audits. - [EXTERNAL_DOWNLOADS]: The skill relies on the
unlighthouseNPM package and includes an installation script (install.sh) to set up the environment. - [DYNAMIC_EXECUTION]: The skill dynamically generates a configuration file (
unlighthouse.config.mjs) to manage parameters like route caps and timeouts, which is the documented method for configuring the Unlighthouse CLI. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external websites during the audit process.
- Ingestion points: External URLs provided by the user for Lighthouse auditing.
- Boundary markers: The skill uses a wrapper script that performs a 'url_safety pre-flight' to validate inputs.
- Capability inventory: Executes subprocesses for auditing and reads generated JSON result files.
- Sanitization: Audit data is normalized into structured JSON (
ci-result.json) before being returned to the agent context.
Audit Metadata