The Agent Skills Directory

[SAFE]: The skill acts as a set of quality control guidelines and manual checklists for developers. It does not perform any automated actions, network requests, or sensitive file access.
[COMMAND_EXECUTION]: The skill suggests the use of standard, non-destructive Unix utilities (ls, head, wc) for local file inspection. These commands are used for structural verification (checking file presence, reading headers, and counting lines) and do not pose a security risk in this context.
[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for processing external data (reviewing other skills). While it lacks explicit boundary markers or sanitization, the intended capabilities are limited to basic file metadata inspection without dynamic code execution or sensitive data sinks.
Ingestion points: Reads files located in skills/[skill-name]/.
Boundary markers: Absent.
Capability inventory: Basic file system inspection via ls, head, and wc.
Sanitization: Absent.
[CREDENTIALS_UNSAFE]: The skill actively encourages safe credential management by instructing users to use .env.example files and ensuring .env files are included in .gitignore.

review-skill