The Agent Skills Directory

[SAFE]: The skill provides a comprehensive workflow for project planning with multiple human-in-the-loop approval stages using structured UI tools to ensure user intent is accurately captured.
[SAFE]: Analysis of the codebase is restricted to identifying architectural patterns and project structure to improve the quality of generated specifications, with no evidence of unauthorized data access.
[SAFE]: No malicious command execution, data exfiltration, obfuscation, or credential theft patterns were found within the skill instructions or reference materials.
[SAFE]: The execution handoff logic recommends the use of platform-native features and separate sessions for implementation, which promotes context isolation and reduces the risk of unintended actions.
[SAFE]: Indirect Prompt Injection Surface: The skill ingests untrusted user input (brain dumps). This surface is well-mitigated by an analytical extraction phase and mandatory user verification of the resulting contract before any code-related artifacts are generated.

ideation