Therapy Mode

Before installing or enabling this skill, consider the following: - Path mismatch: the Python CLI writes to /Users/sethrose/.clawdbot/workspace/therapy-notes (hardcoded). Verify and update the path to a correct, intended workspace directory; otherwise the skill may fail or write sensitive data to someone else's path. - Sensitive data at rest: the skill will persist session notes (potentially PHI). Confirm where files will be stored, whether they are encrypted, and who/what processes can read them. - Destructive operations: the CLI supports hard delete (permanent removal) and archive/restore via filesystem moves. Ensure you understand and control who can invoke these commands (agents with shell access could delete data). - Broad/inexact instructions: SKILL.md asks the agent to 'review entire session' and to 'connect to prior therapy history if available' without specifying safe search scope; that could cause reading unrelated files. Limit the agent's file access or explicitly configure which directories are permitted. - Clinical & safety concerns: the SKILL.md contains clinical directives (MUST complete case formulations, crisis protocols). This is not a substitute for trained clinicians. If you plan to use it clinically, ensure compliance with privacy laws and professional oversight. Recommendations: update the code to use a workspace-relative path (e.g., from an explicit WORKSPACE env var), add explicit configuration for storage location and retention, enable encryption or access controls on note files, restrict execution so only trusted users/agents can run the CLI, and review the SKILL.md for any vague instructions that could cause the agent to access unrelated data. If you cannot verify or modify the hardcoded path and data-handling behavior, treat the skill as risky for storing sensitive therapy notes.