clean-code

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's Agent → Script Mapping explicitly includes commands that run tools against arbitrary URLs (e.g., performance-profiling/scripts/lighthouse_audit.py and webapp-testing/scripts/playwright_runner.py ), and the Script Output Handling sub-skill requires the agent to READ → SUMMARIZE → ASK and then act on those results, meaning untrusted public webpages can be ingested and influence agent decisions.