The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the ls docs/ command to retrieve a directory listing and determine documentation categories.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by incorporating untrusted data from conversation history and codebase files into documentation content and the AGENTS.md index. Ingestion points: Previous user messages and the current codebase. Boundary markers: None identified; there are no delimiters to isolate processed data. Capability inventory: Directory listing, file creation, and file modification. Sanitization: No input validation or escaping is performed on the ingested content.

doc-create