A seven-phase coordinator that validates optimization plans before execution by combining mandatory multi-lane research (official docs, MCP Ref, Context7, web best practices) with parallel agent evidence and local feasibility checks. It loads optimization context from `.hex-skills/optimization/{slug}/context.md`, launches external agents for review, merges findings at a sync barrier, then runs a two-stage refinement process with specific advisor perspectives before issuing a GO/NO_GO verdict. The workflow is heavily structured with phase contracts, mandatory todo items, and explicit self-checks. Useful when you need a reproducible validation pipeline that won't approve plans without source-backed evidence, though the rigid phase progression and evaluation runtime dependency mean this is overkill for quick hypothesis testing.
npx -y skills add levnikolaevich/claude-code-skills --skill ln-813-optimization-plan-validator --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
mattpocock/skills