Coordinate multi-stage software change workflows with staged verification via the megaplan CLI.
npx skills add https://github.com/peteromallet/arnold --skill megaplan
Please help me install this Agent Skill. Skill: megaplan Source: https://github.com/peteromallet/arnold/tree/main/megaplan/data/_codex_skills/megaplan Command: npx skills add https://github.com/peteromallet/arnold --skill megaplan
Megaplan turns brittle, single-pass AI attempts into structured, multi-stage planning and execution with explicit checkpoints that reduce skipped steps and rubber-stamping.
Use megaplan to generate an execution-ready plan and code change for your project by running the verified launcher and then calling megaplan init with your idea and project directory.
High-intent search queries and answers about installing and using this skill.
Robust multi-phase planning workflows for AI code generation use staged verification across prep, plan, critique, gate, revise, finalize, execute, and review phases. This enforces explicit checkpoints that reduce skipped steps and rubber-stamping in software execution.
Per-step model routing allocates cheaper AI models to routine code components and reserves premium adjudication for genuinely hard logic. This model orchestration applies staged verification to balance cost and robustness across software execution workflows.
Execution-ready plans for software refactoring and migrations are generated by running a verified launcher and initializing the CLI with your project directory and idea. The workflow loop semantics enforce structured planning across all execution stages.
Metaplan mode generates single-artefact design and proposal documents instead of source-code diffs. This document-first mode supports architecture artifacts and optionally imports prior Settled Decisions for design-first software execution.
The megaplan workflow requires using its CLI with a verified launcher resolution check before initialization. It operates with no external dependencies, enforcing mode and output validation across prep, plan, critique, gate, revise, finalize, execute, and review stages.
Single-pass AI code generation fails on complex software changes due to brittleness and rubber-stamping without explicit checkpoints. Structured intelligence coordination workflows solve this by enforcing stage-scoped outputs and validation across critique, gate, and revise phases.
corezoidInitialize and configure the Simulator.Company environment with OAuth authentication and .env storage.
AshenDulsankaReview code changes against coding standards and functional specifications.
wtahPlan local adapter patterns for offline testing of cloud services.
djpTester44Generates minimal code to pass a given failing test case for TDD red-to-green phase.
