The Agent Skills Directory

[SAFE]: The skill is composed of architectural guidelines and workflow descriptions in Markdown format. It lacks any scripts, binary files, or configuration that could execute code or perform network operations.
[NO_CODE]: There are no programming language scripts (Python, JavaScript, etc.) or shell commands provided within the skill's file.
[INDIRECT_PROMPT_INJECTION]: The skill involves processing untrusted project documentation and code for review purposes. The workflow is designed for analysis and reporting within the agent's context, and it does not define any actions that would execute or exfiltrate the processed data.

architecture-review