Orchestrate multi-agent Fable 5 code builds with planning and pre-commit review.
npx skills add https://github.com/richkuo/rk-skills --skill fable-advisor
Please help me install this Agent Skill. Skill: fable-advisor Source: https://github.com/richkuo/rk-skills/tree/main/skills/fable-advisor Command: npx skills add https://github.com/richkuo/rk-skills --skill fable-advisor
This skill solves the challenge of maintaining high-quality, reliable code output by separating the execution of tasks from the planning and review phases, ensuring that the agent building the code is not the same entity grading it.
Trigger the fable-advisor skill by typing execute this with a fable advisor followed by your task description or GitHub issue reference.
High-intent search queries and answers about installing and using this skill.
Automated code review and planning for GitHub issue resolution is handled by orchestrating a multi-agent workflow that separates code execution from a persistent planning advisor and a binding pre-commit reviewer. This enforces high-integrity code delivery by ensuring the builder is not the grader.
Separating code execution from planning is achieved by deploying a persistent Fable 5 planning advisor to author plans and consult on checkpoints, while a fresh, independent Fable 5 reviewer issues a final binding verdict before any commit occurs.
Gating automated software development tasks against existing GitHub issues is supported natively. The workflow automatically synchronizes with GitHub, posts implementation plans as comments, and manages the build-and-ship pipeline through rigorous quality gate enforcement.
Pre-commit quality gates are enforced by a separate, independent Fable 5 reviewer subagent that evaluates the built code and issues a binding final verdict before any commit is authorized within the automated GitHub pipeline.
The multi-agent code review workflow requires no external dependencies to function. It operates as a self-contained skill that orchestrates specialized subagents for planning, execution, and review phases directly within your environment.
Using a separate planning advisor and reviewer instead of a single AI agent prevents conflicts of interest where the entity building the code also grades it. This separation maintains high-quality, reliable code output through independent checkpoint consultation and binding verification.
BURHANDEV-ENTERPRISECoordinate GitHub issue tracking with multi-agent AI swarm orchestration.
ClearMeasureLabsOrchestrate GitHub issue implementation via isolated subagents for cloning, branching, and PR submission.
incubateur-ademeIdentify code quality issues and security risks in diffs.
cloudfieldczOrchestrate tasks with fresh subagents and automated two-stage reviews within one session.
