The Agent Skills Directory

[SAFE]: The skill is primarily composed of instructional markdown files and checklists that promote high engineering standards and performance measurement. It contains no executable scripts or malicious instructions.
[COMMAND_EXECUTION]: The instructions recommend several standard command-line tools for performance and accessibility auditing (e.g., npx lighthouse, npx bundlesize, npx pa11y). These are well-known developer utilities used for synthetic testing and CI/CD validation.
[INDIRECT_PROMPT_INJECTION]: The skill establishes an ingestion surface by instructing the agent to analyze external data sources such as application performance logs, RUM metrics, and database query logs. This is documented as a risk factor; however, the skill provides mitigating guidance by emphasizing measurement against fixed baselines and requiring human review in the 'Verify' step of the workflow.
[SECURITY_PRACTICES]: A dedicated security checklist is included that explicitly warns against common vulnerabilities, such as hardcoded secrets, lack of input validation, and the unsafe handling of LLM outputs (eval/shell injection), demonstrating a strong security posture.

performance-optimization