The Agent Skills Directory

[SAFE]: The skill performs interactive user interviews and generates documentation based on user responses. No malicious patterns such as prompt injection, obfuscation, or persistence mechanisms were detected.
[COMMAND_EXECUTION]: The skill instructs the agent to save files to the workspace/prep/ and ~/Development/artifacts/ directories. These operations are essential for its stated purpose of creating flow inventory and specification documents. The use of the /visualize tool is consistent with the goal of generating flow diagrams.
[DATA_EXFILTRATION]: No network exfiltration or sensitive file access patterns were identified. The network operations are limited to internal tools for visualization, and file operations target standard project-specific artifact paths.
[INDIRECT_PROMPT_INJECTION]: The skill uses interactive questioning (AskUserQuestion) to gather data from the user. While this represents a data ingestion surface, the context is limited to documenting project architecture. The agent acts as an interviewer, following a rigid 9-category framework, which reduces the surface area for unexpected behavior.

Flow Discovery