What is a 'bead' in this context?

A bead is a granular, atomic task unit decomposed from a technical design specifically formatted for implementation by an autonomous agent.

How does this skill prevent agent confusion?

It enforces 'Context Auditing' which ensures no implicit references exist, external APIs have clear contracts, and acceptance criteria are testable and unambiguous.

When is a bead review considered complete?

The review reaches convergence when the schema is valid, zero blockers exist in two consecutive passes, and no regressions are detected.

What does the Review Synthesizer do?

It merges findings from all sub-agents, removes duplicate issues, and classifies findings as new, known-open, regressions, or waived to create a clear verdict.

Why use parallel sub-agents for the review?

Large task lists (BEADS.md) can be thousands of lines long; parallel specialized agents maintain significantly higher accuracy and focus than a single generalist agent.

Bead Review Auditor

Name: Bead Review Auditor
Author: quantfiction

byquantfiction

安全与测试

Performs adversarial QA audits on bead-based task decompositions to ensure reliable autonomous agent execution.

关于

Bead Review is a specialized workflow designed to audit complex task decompositions (beads) before implementation by autonomous agents. By utilizing five parallel sub-agents—covering requirements mapping, atomicity, dependencies, context, and metadata—it identifies potential 'agent confusion' points or wasted work. This skill ensures that every task in a Technical Design Document is accurately represented, properly sized, and logically sequenced, providing a robust ledger of findings that prevents compounding errors during the development phase.

主要功能

Parallel sub-agent architecture for specialized audit coverage
Convergence tracking with a persistent audit ledger to ensure stability
Context and metadata verification for self-sufficient task definitions
0 GitHub stars
Dependency graph validation to prevent circular blocks or missing prerequisites
Automated TDD-to-bead requirement mapping and gap analysis

使用场景

Auditing complex implementation plans before handing them off to an AI agent
Identifying circular dependencies or hidden work in large-scale project migrations
Validating that a Technical Design Document has been fully decomposed into actionable tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add quantfiction/dotfiles bead-review

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于