ralph

Autonomous feature development is automated by breaking work into user stories and executing them iteratively. The Skill orchestrates fresh AI agent cycles to implement, verify, and integrate features while tracking progress in prd.json and progress.txt.

PRD-driven iterative development uses a product requirements document to guide AI agents through incremental code changes. It maintains persistent state across runs using prd.json for requirements and progress.txt for tracking learnings and completed work.

Yes, jq and droid are required dependencies for running this autonomous feature delivery workflow. These tools support the underlying script components that orchestrate the AI agent iterations and manage the persistent state files.

To start the autonomous feature implementation loop, execute the provided script. This initiates the workflow which generates a PRD, orchestrates per-iteration agent execution for code changes, and commits the results automatically.

Each iteration cycle runs a new AI agent with clean context to implement a single user story. This prevents context bloat and ensures the agent focuses solely on the specific requirements defined in the prd.json state file.

This approach is limited to software projects adopting a PRD-driven iterative workflow. It relies heavily on the droid dependency and requires git history for persistent memory, meaning it may not suit projects without structured requirement files.