The Agent Skills Directory

[SAFE]: The skill consists entirely of markdown documentation and philosophical guidelines for software design. It does not include any scripts, binaries, or tool definitions that could perform actions on the host system.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for reviewing external code or designs, which constitutes a surface for processing untrusted data. However, the skill's core directive to 'bias toward deletion' acts as a natural heuristic against many types of additive malicious logic.
Ingestion points: User-provided source code and design documentation analyzed during refactoring tasks.
Boundary markers: None explicitly defined in the skill instructions to separate code-under-review from agent instructions.
Capability inventory: The skill itself does not define or use any tools (subprocess, network, or file-system operations).
Sanitization: No specific sanitization or filtering instructions are provided for the content being reviewed.
[EXTERNAL_DOWNLOADS]: The skill references several external URLs for further reading. These links point to well-known technical resources (e.g., InfoQ, YouTube, Python PEPs) and established developer blogs. These are provided for informational purposes and do not involve automated downloads or execution of remote code.

reducing-entropy