The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local shell commands and git operations (add, commit, reflog) to manage plan files within the workspace. These commands are standard for the intended workflow and are restricted to the local project environment.
[DATA_EXFILTRATION]: All file operations, including reading the plan template and writing new plan files, are restricted to the local ai-workspace/plans/ directory. No network operations or external data transfer mechanisms were found.
[PROMPT_INJECTION]: The instructions establish a rigid template-based output format and include a 'Threat model selection' step. This design encourages the agent to consider the security context of a task rather than blindly following user-provided task descriptions.

plan