flywheel-fix

Applies targeted code fixes through a single-bead agent workflow with review.

3|1|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/burningportra/agent-flywheel-plugin --skill flywheel-fix-burningportra
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flywheel-fix
Source: https://github.com/burningportra/agent-flywheel-plugin/tree/main/skills/flywheel-fix
Command: npx skills add https://github.com/burningportra/agent-flywheel-plugin --skill flywheel-fix-burningportra

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Running a full multi-phase flywheel pipeline is overkill for small, focused bug fixes. This Skill provides a fast path that applies a targeted fix without the full scan-discover-plan-implement-review cycle, while still keeping tracking, file reservations, and review in place. ## Core Features & Use Cases - Scoped Code Analysis: Uses an Explore subagent to analyze the relevant code and understand the scope before making changes. - Isolated Implementation Agent: Spawns a background general-purpose agent in a git worktree with file reservations to prevent conflicts. - Fresh-Eyes Review: Calls flywheel_review after implementation so the fix gets an independent review pass before commit. - Use Case: You spot a null-pointer bug in one module. Describe the fix, and the Skill creates a bead, delegates the change to an isolated agent, reviews the result, and asks for your confirmation before closing. ## Quick Start Ask the agent to fix the null check in the payment validation module using the flywheel fix fast path.

Frequently Asked Questions about flywheel-fix

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I apply a quick bug fix without a full agent pipeline?

Use the flywheel-fix fast path: describe the fix, and it creates a single bead, analyzes the relevant code with an Explore agent, and delegates the change to one background implementation agent. It skips the full scan-discover-plan phases.

How does the fix agent avoid file conflicts with other agents?

The implementation agent reserves the files it will edit via file_reservation_paths before making changes and releases them when done. It also runs in an isolated git worktree so its changes do not interfere with other work.

When should I use flywheel-fix instead of the full flywheel?

Use it for small, focused changes such as single bug fixes where the full six-phase loop would be unnecessary overhead. Larger features or multi-file refactors should go through the complete flywheel workflow.

What happens if the fix agent goes idle or stops responding?

The workflow nudges the agent with a status request message if it goes idle without reporting. The saved task ID also allows stopping the agent with TaskStop if it cannot recover.

Does the fix get reviewed before committing?

Yes. After implementation completes, the Skill calls flywheel_review for a fresh-eyes review, shows you the results, and only closes the bead after you confirm the changes look good to commit.