The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local scripts provided within the skill package (scripts/set-doc-triplet.sh and scripts/check-doc-triplets.py) to manage file symlinks and audit documentation triplets.
[DATA_EXPOSURE]: The fact-verification guide instructions lead the agent to search the codebase for sensitive identifiers such as constants and environment variables (ENV_VAR) using ripgrep. While this is intended for documentation accuracy, it involves accessing potentially sensitive configuration names.
[INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it instructs the agent to read and process untrusted repository data (code, configuration, and existing documentation) to build its 'fact inventory' without explicit sanitization or boundary markers.

readme-agents-writer