Plannable

⭐ Stars are appreciated!

Spec & Command driven planning for AI coding agents.

Plannable turns a product idea into a short, human-readable MASTER_PLAN.md plus compressed, agent-readable part files (PlannablePlan .ai.md format). Agents load one part at a time — each part carries enough context about the whole masterplan to work alone — and nothing gets checked off without recorded evidence.

Why Plannable (vs spec-kit)

spec-kit is specification-first; Plannable is execution-first.

Pick spec-kit for heavyweight specification governance. Pick Plannable when you want an agent to execute a plan without context rot or unearned checkmarks.

When to Use It

Long agent sessions fail in predictable ways: the plan is too big to keep in context, the agent drifts from it mid-implementation, and by the end it has checked off work it never finished. Plannable is planning built to survive execution — hence the name. The plan is split into parts small enough to actually hold in context, each part carries the whole arc with it, and no checkbox moves without proof.

Reach for it when you're starting from a product idea and want a working first plan in seconds instead of a blank page. When a feature spans multiple agent sessions and each session has to pick up exactly where the last one stopped. When you come back to review an agent's work and want the checkmarks to mean something — every completed part points to evidence naming the files, checks, and notes that prove it.

Plan basic safeguards without changing the workflow

Unreleased source improvement: create now includes secret handling and changed-behavior checks; projects declaring dependencies also receive dependency/configuration review. During the existing enrichment step, the agent inspects facts and adds only applicable…