The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The a11y-auditor agent is designed to ingest and analyze external code files from a user's repository, which constitutes an indirect prompt injection surface.
Ingestion points: Source code files including HTML, JSX, TSX, JS, TS, and CSS are read using the Read, Glob, and Grep tools as specified in agents/a11y-auditor.md.
Boundary markers: The agent is provided with a structured reporting template, but lacks explicit instructions to ignore natural language instructions that might be embedded in the code comments or content being audited.
Capability inventory: The agent's capabilities are restricted to read-only operations via Read, Glob, and Grep. It does not have network access or file-write permissions.
Sanitization: The skill does not instruct the agent to sanitize or escape content retrieved from audited files before processing or including it in the final report.

accessibility