The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses firecrawl:firecrawl_scrape and firecrawl:firecrawl_search to ingest external performance documentation and optimization guides into the agent context. This introduces a surface where malicious instructions embedded in third-party websites could attempt to influence the agent's behavior.
Ingestion points: firecrawl:firecrawl_search and firecrawl:firecrawl_scrape in SKILL.md.
Boundary markers: None provided in the instructions to separate external content from internal logic.
Capability inventory: The skill has the capability to run local shell commands (npm, npx), query Sentry for events, and execute SQL EXPLAIN commands.
Sanitization: No explicit sanitization or filtering of external content is mentioned before it is processed by the agent.
[REMOTE_CODE_EXECUTION]: The skill suggests using npx source-map-explorer. While this is a well-known and standard development tool, npx by definition downloads and executes code from a remote registry. This is expected behavior for a performance audit tool and targets a well-known service (npm registry).

audit-performance