Does the fix skill edit code files directly?

No, the fix skill is strictly an orchestrator. It dispatches specialized subagents to perform the actual code edits, ensuring a clean separation between task management and implementation.

Why does the skill require a NEXT_ITER_TASK_ID?

The task ID is used to programmatically block the next review iteration. This ensures that the review process cannot proceed until all identified critical and major issues have been resolved.

What happens if a review finding is a false positive?

During the process, the skill displays a findings table where issues can be marked as 'SKIP'. Trivial issues or false positives are noted in the summary but do not trigger subagent dispatch.

Can this skill fix multiple issues simultaneously?

To maintain code stability and prevent conflicts, the fix skill follows 'Iron Rules' that mandate sequential execution, processing one issue at a time.

Review Loop Fix Agent

Name: Review Loop Fix Agent
Author: onsails

byonsails

•

Collaboration Tools

Orchestrates subagents to automatically resolve code review findings through sequential, task-based remediation.

The fix skill is a specialized component of the review-loop plugin designed to automate the transition from code review to code resolution. Operating under a strict 'dispatch-only' model, this skill prevents the primary agent from making direct edits, instead parsing review findings and delegating specific tasks to specialized subagents. It maintains project integrity by creating a structured task list, blocking subsequent review iterations until fixes are verified, and providing a clear audit trail of fixed versus skipped issues. This is ideal for maintaining high code quality standards without manual intervention.

Key Features

01Automatic task creation and blocking management for review iterations

02Automated subagent orchestration for hands-off code remediation

039 GitHub stars

04Comprehensive summary reporting of fix status and iteration readiness

05Structured severity-based prioritization (Critical, Major, Minor)

06Sequential execution logic to prevent merge conflicts and race conditions

Use Cases

01Enforcing 'fail fast' standards by ensuring all major issues are resolved before the next test cycle

02Automating the remediation of security vulnerabilities found during AI-driven audits

03Managing complex refactoring tasks across multiple files during iterative reviews

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add onsails/cc fix

For use in Claude.ai and ChatGPT

Download Skill