The Agent Skills Directory

[SAFE]: The skill implements a state-machine based workflow for project planning, strictly enforced through human-gated steps that require explicit approval to advance phases.
[SAFE]: No network activity, remote downloads, or unauthorized file access was detected. All file operations are restricted to a user-provided or session-based local directory.
[SAFE]: The skill interacts with companion skills (such as grill-me, to-prd, and to-issues) but requires them to be present in pre-defined local directories, preventing arbitrary skill execution or remote fetching of unknown skills.
[SAFE]: The Python implementation uses standard library modules for local file manipulation and environment variable reading. No dynamic execution patterns (eval, exec) or subprocess calls to external binaries were found.
[SAFE]: The instructions and code do not contain any obfuscation, persistence mechanisms, or privilege escalation attempts.

orchestration-planner