The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches documentation (llms.txt) and individual markdown pages from the project's official website (mikeyobrien.github.io). These downloads are restricted to documentation files and are stored in a local cache (~/.cache/ralph-docs).
[COMMAND_EXECUTION]: The skill instructions utilize curl for fetching remote text files and mkdir for creating cache directories. These are standard operations for the skill's documentation-retrieval purpose.
[DATA_EXFILTRATION]: No patterns for accessing sensitive user data or exfiltrating information to third-party servers were found. All network operations are directed at the author's own project domain.
[PROMPT_INJECTION]: No evidence of prompt injection, safety bypass, or instructions to ignore system guidelines was detected.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external content (documentation pages). While this represents a potential attack surface for indirect injection if the source were compromised, the risk is minimized by the use of authoritative project-owned URLs and the skill's specific focus on information retrieval.

ralph-docs