The Agent Skills Directory

[SAFE]: The skill defines a behavioral flow for interviewing users. It interacts with the local environment using the @explore tool for context and writes results to the .omc/specs/ directory. No malicious intent or vulnerabilities were found.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: The tool uses @explore to analyze the codebase, which is a standard procedure for context-aware agents. There is no evidence of data being sent to external or untrusted domains.\n- [COMMAND_EXECUTION]: The skill references platform tools such as ralplan, omg-autopilot, and ralph for the execution phase. These are internal component references and do not represent arbitrary or malicious command execution.

deep-interview