The Agent Skills Directory

[SAFE]: The skill performs documentation formatting tasks using a defined set of local tools including file reading, editing, and restricted shell commands. It includes a local helper script to process configuration files at the repository root.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from existing README files, creating a theoretical ingestion surface for indirect prompt injection.
Ingestion points: The skill reads the entire content of the existing README.md file in SKILL.md (Step 1).
Boundary markers: No explicit boundary markers or isolation instructions are provided for the input text.
Capability inventory: The skill has access to file system tools (Read, Write, Edit, Glob, Grep) and restricted Bash capabilities (jq, wc).
Sanitization: The skill instructions focus on verbatim structural transformations (e.g., wrapping content in <details> blocks), which limits the interpreter's interaction with the text and reduces the risk of following embedded instructions.

compact-readme