rune-fix

Applies diagnosed code fixes and verifies changes with tests, lint, and type checks.

1|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/dangvu008/VietTruyen --skill rune-fix-dangvu008
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: rune-fix
Source: https://github.com/dangvu008/VietTruyen/tree/main/.agents/skills/rune-fix
Command: npx skills add https://github.com/dangvu008/VietTruyen --skill rune-fix-dangvu008

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Applying bug fixes without a clear diagnosis or verification often introduces new regressions, runaway fix loops, and broken tests. This Skill enforces a disciplined locate-change-verify workflow so every code change is grounded in a root-cause diagnosis and validated before being reported. ## Core Features & Use Cases - Diagnosis-Gated Fixing: Requires a debug report or clear error before touching code, and classifies errors into a recovery policy matrix (auto-fix, retry, prompt user, abort, or re-diagnose). - Quality Decay Self-Regulation: Tracks a WTF-likelihood score across repeated fixes and stops when continued changes risk making things worse, with a hard cap of 30 fixes per session. - Defense-in-Depth Hardening: Adds validation at entry point, business logic, environment, and instrumentation layers so fixed bugs become structurally impossible. - Use Case: After a debug session identifies a null-pointer root cause in an auth module, invoke this Skill to apply the minimal fix, run the failing test then the full suite, verify types and lint, and emit a structured Fix Report with a DONE or DONE_WITH_CONCERNS status. ## Quick Start Ask the agent to apply the fix from the latest debug report and verify it with the project's test suite and type checker.

Frequently Asked Questions about rune-fix

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

FAQPage Schema
How do I apply a bug fix without breaking existing tests?

Apply the minimal change targeting only the diagnosed root cause, then run the specific failing test first followed by the full suite. Never modify test files to make them pass; fix the implementation code instead.

What should I do before fixing a bug in code?

Obtain a diagnosis first, either from a debug report or a clear error description. Classify the error type (input, timeout, logic, dependency, policy) to choose the right recovery strategy before editing any files.

When should I stop applying fixes and re-diagnose?

Stop after 3 failed fix attempts on the same issue and re-run root cause analysis. Also stop when fixes touch files outside the original diagnosis scope or when each fix creates a new failure elsewhere, which signals a structural problem.

Can this fix workflow handle TypeScript type errors?

Yes, after editing .ts or .tsx files it runs TypeScript compilation to confirm no type errors remain, alongside lint checks. It also enforces conventions like avoiding the any type during edits.

Why do repeated automated fixes sometimes make code worse?

Each additional fix adds regression risk, especially when blast radius expands beyond three files or fixes are applied without running tests between them. Quality decay scoring detects this pattern and halts the loop before damage accumulates.