The Agent Skills Directory

[SAFE]: The skill implements a comprehensive methodology for software planning and peer review, focusing on correctness, concurrency, and reliability.
[SAFE]: Bundled analysis scripts, such as check_plan.py and audit_plan_claims.py, perform static validation of plan structure and verify that referenced repository artifacts (files, commands, and scripts) exist locally.
[SAFE]: The audit_plan_claims.py script invokes just --summary via a subprocess to validate task runner recipes. This is a standard developer tool operation restricted to the local project context.
[SAFE]: The instructions include explicit safety constraints, directing the agent that planning and review activities must not execute commits, migrations, or deployments against real systems.

plan-implementation