The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill's behavior is consistent with its stated purpose of managing project documentation and session state.\n- [CREDENTIALS_UNSAFE]: The skill incorporates a proactive security measure by explicitly instructing the agent to keep secrets out of generated files, recording only that a constraint exists rather than its sensitive value.\n- [INDIRECT_PROMPT_INJECTION]: While the skill ingests data from the repository's source tree and build configurations, the risk of indirect injection is mitigated by the highly structured output templates and the requirement for all facts to be traced to specific paths, commands, or versions.

plexy-save