What problem does it solve?
Running a full AI-DLC lifecycle manually requires constant human hand-holding at every stage: elaborating ideas, drafting proposals, reviewing documents, dispatching tasks, and verifying results. This Skill removes that overhead by driving the entire Idea -> Proposal -> Execute -> Verify pipeline autonomously from a single natural-language prompt.
Core Features & Use Cases
- End-to-End Automation: Creates the project and idea, self-elaborates requirements, drafts the proposal with documents and task DAGs, then executes tasks in dependency-ordered waves using parallel subagents.
- Adversarial Review Loops: Spawns proposal-reviewer, task-reviewer, and code-reviewer subagents with bounded retry rounds, escalating to a human only when review rounds are exhausted.
- Pre-Ship Code-Review Gateway: Reviews the aggregate change for the whole idea before writing the mandatory completion report, looping fix tasks through the pipeline on FAIL.
- Use Case: Describe a feature in one prompt, walk away, and return to a verified, code-reviewed implementation with a full audit trail of elaboration Q&A, reviewer verdicts, and work reports in Chorus.
Quick Start
Run /chorus-yolo followed by a detailed natural-language description of the feature you want built, optionally adding --project with an existing project UUID.