The Agent Skills Directory

[SAFE]: The skill provides a standard methodology for performance analysis, focusing on Core Web Vitals and stack profiling without containing any executable scripts or binary files.
[NO_CODE]: The skill consists entirely of instructional markdown and does not distribute, download, or execute any external code or scripts.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a theoretical attack surface through data ingestion but remains within safe parameters for its use case. (1) Ingestion points: Reads Lighthouse reports, database logs, bundle analysis output, and application source code. (2) Boundary markers: The instructions do not specify explicit delimiters for external content analysis. (3) Capability inventory: The skill's functionality is limited to diagnostic analysis and reporting; it does not contain capabilities for file writing, network exfiltration, or shell command execution. (4) Sanitization: SKILL.md explicitly states that user-provided values are treated as read-only targets and are not used to construct commands or file paths.

performance-audit