The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were detected in the skill definition. The skill is strictly limited to documentation generation based on codebase analysis and does not utilize risky system commands or network tools.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data (repository files and trees) as part of its core functionality. However, it explicitly instructs the agent to treat this data as evidence for documentation purposes only and not as a source of authority to execute commands. The risk is minimized by the restricted output format (JSON documentation schema) and the absence of high-privilege tool usage.

wiki-architect