The Agent Skills Directory

[SAFE]: The skill provides legitimate instructions for web performance auditing using industry-standard tools like Google Lighthouse.
[COMMAND_EXECUTION]: Includes common shell commands for running performance audits and managing CI workflows.
[EXTERNAL_DOWNLOADS]: Recommends installing standard, well-known packages such as lighthouse and @lhci/cli from the npm registry, as well as the standard treosh/lighthouse-ci-action GitHub Action.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes data from external tool outputs.
Ingestion points: report.json content read via fs.readFileSync in SKILL.md.
Boundary markers: None; the skill assumes the data follows the Lighthouse JSON schema.
Capability inventory: Shell command execution for audits and file system read access.
Sanitization: None; it extracts numeric metrics directly from the JSON object.

perf-lighthouse