sdcorejs-repair-loop

Verify, fix, and re-verify review findings until Critical and Important items resolve.

2|Updated Apr 18, 2026
One-click install
npx skills add https://github.com/sdcorejs/sdcorejs-agent --skill sdcorejs-repair-loop
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sdcorejs-repair-loop
Source: https://github.com/sdcorejs/sdcorejs-agent/tree/main/plugin/skills/sdcorejs-repair-loop
Command: npx skills add https://github.com/sdcorejs/sdcorejs-agent --skill sdcorejs-repair-loop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

A review or verification that produces findings without an effective closed-loop fix cycle wastes time and creates churn, because issues must be validated, fixed, and re-checked against the same source of truth.

Core Features & Use Cases

  • Finding verification first: Verifies each finding is genuine before touching code to avoid mis-scoped or redundant changes.
  • Tiered fix strategy: Auto-applies mechanical issues, shows diffs for confirm-then-apply items, and defers architectural/user-decision items.
  • Source-aligned re-verification: Re-runs the correct re-check flow based on findings origin (review-code, verify-before-done, linter, or manual), then iterates until Critical + Important are resolved or deferred.
  • Loop convergence guardrails: Hard-caps at 3 iterations and escalates to user when findings do not converge.

Quick Start

Invoke sdcorejs-repair-loop after sdcorejs-review-code-<track> or verify-before-done outputs findings by providing the findings list and the correct source value so the re-verify step runs for the same context.

Frequently Asked Questions about sdcorejs-repair-loop

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

FAQPage Schema
How do I automate fixing lint and typecheck errors after code review?

Automating lint and typecheck fixes requires a closed-loop repair process that verifies findings, applies batch edits, and re-runs checks until Critical and Important issues are resolved. This prevents manual re-verification churn by iterating up to three times.

What is a fix loop for code review findings and how does it work?

A fix loop for code review findings is an automated cycle that verifies each issue, categorizes it into fix tiers, applies mechanical fixes, and re-verifies the source. It converges by resolving or deferring Critical and Important items within three iterations.

How do I handle code review findings that keep failing re-verification?

When code review findings keep failing re-verification after three automated iteration passes, the repair loop escalates the non-convergent issues to the user. This guardrail prevents infinite loops and prompts manual intervention for unresolved items.

Can I batch edit code to resolve linter and typecheck findings automatically?

Yes, you can batch edit code to resolve linter and typecheck findings automatically by applying a tiered fix strategy. The system auto-applies mechanical issues, shows diffs for confirmation items, and defers architectural decisions.

Does the repair automation workflow support manual human-reported issues?

Repair automation supports manual human-reported issues by selecting source-aligned re-verification commands. It processes findings from manual reports, linters, or review tracks, categorizes them, and iterates fixes until convergence.

Why verify code review findings before applying automated fixes?

Verifying code review findings before applying automated fixes ensures each issue is genuine, preventing mis-scoped or redundant changes. This validation step avoids unnecessary code churn and maintains source integrity during the repair loop.