Zero Debt

AI coding agents can write code. Proving the work is complete is the harder part. Decapod is a repo-native governance kernel for builders using AI coding agents across tools such as Cursor, Claude Code, Codex, Antigravity, and Grok. It helps you turn a natural-language task into bounded, durable, proof-backed work by keeping specifications, execution state, validation, and evidence inside the repository. Key features: • Explicit intent – turns vague requests into versioned specifications and todos • Scoped context – creates provenance-backed context capsules for the relevant code and docs • Concurrent-work controls – supports exclusive task claims and isolated workspaces for agents • Proof-gated completion – requires deterministic verification before work can reach VERIFIED • Harness flexibility – governs agent work without replacing the model, agent, or harness you use It’s open-source (MIT license). Link in the reply 👇