octocode-plan

Software research findings are converted into actionable implementation plans by orchestrating research-driven planning that produces a traceable plan.md with clear dependencies, risk flags, and references to the original research.

Evidence-based planning for multi-repo software projects coordinates local and external research tasks across single or multiple repositories, producing versioned session contexts and a plan.md that references consolidated research findings.

AI agents are coordinated for research tasks through a phase-driven workflow that automatically allocates local and external research to specialized skills, consolidating results before generating an approved step-by-step implementation plan.

Explicit approval gates are enforced between the Understand, Research, Plan, Implement, and Verify phases, ensuring phase-driven workflow requires explicit approvals before execution of any implementation step begins.

Phase-driven planning supports traceability by producing an evidence-backed plan.md that explicitly references research findings, capturing decisions with clear dependencies and risk flags before any coding starts.

Planning a feature across multiple software modules is best handled by starting a named planning session that collects context, discovers research findings via delegated tasks, and generates a detailed plan.md with approved steps and multi-repo coordination.