Coordinate multi-agent planning, implementation, and validation for complex coding tasks.
npx skills add https://github.com/tiancaiamao/ai --skill pge
Please help me install this Agent Skill. Skill: pge Source: https://github.com/tiancaiamao/ai/tree/main/skills/pge Command: npx skills add https://github.com/tiancaiamao/ai --skill pge
This Skill solves the challenge of delivering complex coding work that needs clear planning, independent implementation, and objective verification without relying on a single self-reviewing agent.
Use the pge skill to turn this complex coding request into a spec, split it into tasks, and iterate implementation and independent validation until the acceptance criteria pass.
High-intent search queries and answers about installing and using this skill.
Multi-agent orchestration for complex coding tasks coordinates independent planner, generator, and evaluator roles to decompose specs, implement code, and verify acceptance criteria through structured handoff files and isolated validation loops.
Spec-driven decomposition for multi-file feature work uses structured task files and acceptance criteria to split work into isolated generator and evaluator agents, ensuring precise, auditable handoffs and objective verification rather than relying on self-review.
Yes, independent evaluator agents support repeated bug-fix loops by enforcing up to three validation rounds with build and review gates, iterating implementation and verification until the code meets stated requirements or stops to report back.
Multi-agent orchestration works for architecture changes by applying isolated generator and evaluator agents to handle multi-file refactors, using structured handoffs to maintain objective evaluation and independent implementation throughout the validation rounds.
Spec-driven execution requires structured handoff files for large features to keep work precise and auditable, preventing self-reviewing agent bias by ensuring the generator and evaluator agents operate independently based on the acceptance criteria.
A single self-reviewing agent lacks independent evaluation for complex product changes, whereas multi-agent orchestration enforces spec-driven decomposition and isolated validation loops to objectively verify acceptance criteria and converge through repeated implementation.
lucasbenevinutoReads a search file and generates a spec file with implementation strategy and acceptance criteria.
Fr-e-dApply minimal fixes to QA failures and iterate until quality gates pass.
bostonaholicSeparate generator and evaluator roles to review code against security and quality standards.
jpoutrinOrchestrates parallel Codex and Gemini agents for PRD-driven software development.
