New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Narrow Debug Fix

by lge-ros2 in lge-ros2/cloisim

Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation.

MITAuto-check passedDevelopment

Install Narrow Debug Fix

skills CLI
$ npx skills add lge-ros2/cloisim --skill narrow-debug-fix -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install lge-ros2/cloisim narrow-debug-fix --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/lge-ros2/cloisim.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.github/skills/narrow-debug-fix .claude/skills/narrow-debug-fix && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
narrow-debug-fix
GitHub stars
176
Token cost
~1.2k tokens
SKILL.md length
663 words
Files
1
Skills in repo
18
Repo updated
First seen
Licence
MIT

At a glance

Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation.

  • Works in 7 steps: Start from a Concrete Anchor → Read Only Enough Local Context → State the Working Hypothesis → …
  • : debugging a failing test
  • SKILL.md covers When to Use, Outcome, Procedure and Repo Guardrails, plus 1 more section
  • Calls git

What it does

Narrow Debug Fix is an agent skill from lge-ros2/cloisim. Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation. Use when: debugging a failing test, regression, runtime error, compiler error, wrong behavior, or a review comment asking for a narrow fix.

Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development, covering Failing and flaky tests and Debugging. The repository describes itself as: Unity 6 based multi-robot simulator for ROS 2, SDFormat/SDF, LiDAR, camera, depth, IMU, GPS, and large-scale robotics simulation. The licence is MIT.

When your agent uses it

  • : debugging a failing test
  • A review comment asking for a narrow fix

Example prompts

  • “/narrow-debug-fix”

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. Start from a Concrete Anchor
  2. Read Only Enough Local Context
  3. State the Working Hypothesis
  4. Choose the Cheapest Discriminating Check
  5. Make the Smallest Grounded Edit
  6. Validate Immediately After the First Substantive Edit
  7. Close Out

What it can do on your machine

Read from SKILL.md and the folder at commit f4c1f0b. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Narrow Debug Fix loads about 1.2k tokens when it runs. Until then it costs about 84 tokens; SKILL.md has 663 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~84
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from lge-ros2/cloisim at commit f4c1f0b, republished under its MIT licence (© lge-ros2). 663 words, ~1,224 tokens.

Download SKILL.mdSave it as .claude/skills/narrow-debug-fix/SKILL.md (or your agent's skills folder).
name
narrow-debug-fix
description
Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation. Use when: debugging a failing test, regression, runtime error, compiler error, wrong behavior, or a review comment asking for a narrow fix.
argument-hint
Describe the failing test, command, error, symbol, or behavior.

Narrow Debug and Fix

Use this workflow to keep bug fixing local, testable, and falsifiable. It is optimized for this repository's style: narrow code reads, small edits, immediate validation, and no unrelated cleanup.

When to Use

  • Failing tests with a known test name, command, or log line
  • Regressions after a recent code change
  • Compiler, lint, or type errors confined to a touched slice
  • Runtime exceptions with a nearby owner in code
  • Incorrect behavior with a concrete file, symbol, or command anchor
  • Review comments asking for a narrowly scoped repair

Outcome

By the end of the workflow you should have:

  • One identified controlling code path
  • One falsifiable explanation for the failure
  • One small edit that addresses the root cause or exposes the missing piece
  • One focused validation result
  • One short summary of residual risk or uncovered cases

Procedure

1. Start from a Concrete Anchor

Prefer one of these anchors:

  • A file named by the user
  • A failing test name
  • A failing command
  • Exact error text
  • A nearby symbol or call site

If the first file mainly forwards, registers, or wires behavior, step one hop to the code that directly computes, mutates, or controls it.

2. Read Only Enough Local Context

Gather just enough evidence to answer:

  • What code path currently decides the behavior?
  • What is the smallest nearby abstraction that owns that decision?
  • What cheap check could prove this theory wrong?

Stop once you can name:

  • One falsifiable local hypothesis
  • One cheap discriminating check
  • One smallest plausible edit
3. State the Working Hypothesis

Write the hypothesis as a claim that can fail. Examples:

  • This guard drops the request before the plugin registers.
  • This conversion uses the wrong axis order for nested models.
  • This reset path forgets to clear the cached message.

Avoid carrying multiple competing theories unless one nearby read is required to distinguish them.

4. Choose the Cheapest Discriminating Check

Prefer checks in this order:

  1. The failing behavior or command itself
  2. A narrow test covering the touched slice
  3. A slice-scoped compile, lint, or typecheck
  4. A repo script that exercises the same subsystem

git diff is not validation when an executable check exists.

5. Make the Smallest Grounded Edit
  • Fix the root cause rather than the visible symptom when the local evidence supports it.
  • Preserve existing APIs and surrounding style.
  • If confidence is incomplete, make a small reversible probe that exposes the control-flow gap or type mismatch.
  • Do not widen scope to adjacent cleanup during the first edit.
Show full SKILL.md (252 more words)Show less
6. Validate Immediately After the First Substantive Edit

Run the same focused check chosen in step 4 before doing more reading or patching.

Interpret the result like this:

  • Validation passes: only make adjacent follow-up edits that are now clearly required, then rerun the same check.
  • Validation fails but still supports the same hypothesis: repair that same slice and rerun the same check.
  • Validation falsifies the hypothesis: step one nearby hop to the next controlling code path and repeat.
  • Validation is ambiguous: do one nearby disambiguating read, then choose between local repair and a one-hop move.
7. Close Out

Finish with at least one executable post-edit validation when possible.

Report:

  • What changed
  • What check passed or failed
  • Any residual risk, missing coverage, or follow-up worth doing next

Repo Guardrails

For CLOiSim-specific work:

  • Route SDF behavior through the parse -> import -> implement pipeline instead of scene-only patches.
  • Preserve plugin lifecycle contracts: OnAwake() -> OnStart() -> Started -> OnReset() / OnDestroy().
  • Do not rename scene roots: Core, Props, World, Lights, Roads, UI.
  • Use tabs and existing C# brace style in Unity scripts.
  • Favor narrow EditMode or subsystem-scoped checks before broad project validation.

Completion Checklist

  • I started from one concrete failure, file, symbol, or command.
  • I identified the code that directly controls the behavior.
  • I can state one falsifiable local hypothesis.
  • I chose one cheap check that could disprove it.
  • My first edit was the smallest plausible change or probe.
  • I validated immediately after that edit.
  • I avoided unrelated refactors or cleanup.
  • I finished with a short outcome and remaining risk.

© lge-ros2, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .github/skills/narrow-debug-fix of lge-ros2/cloisim.

Open the folder on GitHubat commit f4c1f0b

Compare with similar skills

Narrow Debug Fix next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Narrow Debug Fix compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Narrow Debug Fix this skilllge-ros2/cloisim176—~1.2kAutomated safety check: PassMIT
React Router Bug Fix Workflowremix-run/react-router57k—~1.3kAutomated safety check: PassMIT
Runtime Debugvercel/next.js143k1 repos~618Automated safety check: PassMIT
Root Cause Debuggingjsmastery-pro/skills1.4k—~1.8kAutomated safety check: NotesMIT
Superpowers Systematic Debuggingchristopherarter/superpowers-reasonix102—~2kAutomated safety check: PassMIT
Minimal Code Fixcobusgreyling/loop-engineering11k1 repos~345Automated safety check: NotesMIT

Similar skills

  • React Router Bug Fix Workflow

    remix-run/react-router

    Fixes a React Router bug reported in a GitHub issue end to end: fetching the issue, validating the reproduction, writing a failing test and implementing the fix on a new branch.

    57k GitHub stars~1.3k tokensUpdated 2 days ago
    DevelopmentAuto-check passed

  • Runtime Debug

    vercel/next.js

    Official

    Debug and verification workflow for runtime-bundle and module-resolution regressions.

    143k GitHub starsUsed in 1 repo~618 tokens
    DevelopmentAuto-check passed

  • Root Cause Debugging

    jsmastery-pro/skills

    Runs a reproduce, localize, hypothesize, test, fix and verify loop to find a bug's root cause, applies the minimal fix and hands off a regression test.

    1.4k GitHub stars~1.8k tokensUpdated 2 mo ago
    DevelopmentAuto-check: notes

  • Superpowers Systematic Debugging

    christopherarter/superpowers-reasonix

    Any bug, failing or flaky test, or surprise behavior?. An agent skill from christopherarter/superpowers-reasonix.

    102 GitHub stars~2k tokensUpdated 1 mo ago
    DevelopmentAuto-check passed

  • Minimal Code Fix

    cobusgreyling/loop-engineering

    Makes the smallest code change that fixes one well-scoped problem, such as a CI failure, review comment or typo, without refactoring anything unrelated.

    11k GitHub starsUsed in 1 repo~345 tokens
    DevelopmentAuto-check: notes

  • Fix The Class

    joetawil7/first-pass

    Bug-fix routine that fixes the whole class of bug, not just the reported instance.

    92 GitHub stars~1.5k tokensUpdated yesterday
    DevelopmentAuto-check passed

More from lge-ros2/cloisim

All 18 skills in this repo

  • Add Actor Animation

    lge-ros2/cloisim

    Add or modify SDF actors with skeletal animation and waypoint-based navigation.

    176 GitHub stars~2.4k tokensUpdated 3 days ago
    Auto-check passed

  • Add Compute Shader

    lge-ros2/cloisim

    Add a new compute or surface shader for sensor data processing or visualization.

    176 GitHub stars~2.1k tokensUpdated 3 days ago
    Auto-check passed

  • Add Mesh Import

    lge-ros2/cloisim

    Import 3D mesh files into CLOiSim via the Assimp pipeline, including collision generation and procedural meshes.

    176 GitHub stars~2k tokensUpdated 3 days ago
    Auto-check passed

  • Add Motor Controller

    lge-ros2/cloisim

    Add a new motor or drive controller for robot locomotion. An agent skill from lge-ros2/cloisim.

    176 GitHub stars~1.4k tokensUpdated 3 days ago
    Auto-check passed

  • Add New Plugin

    lge-ros2/cloisim

    Add a new CLOiSimPlugin for bridging a device to the ROS 2 transport layer.

    176 GitHub stars~1.5k tokensUpdated 3 days ago
    Auto-check passed

  • Add New Sensor

    lge-ros2/cloisim

    Add a new sensor device to CLOiSim, including the Device subclass, SDF pipeline wiring, and transport plugin.

    176 GitHub stars~1.6k tokensUpdated 3 days ago
    Auto-check passed

Categories

Questions about Narrow Debug Fix

What does Narrow Debug Fix do?

Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation. Narrow Debug Fix is an agent skill from lge-ros2/cloisim. Debug and fix a bug by starting from one concrete failure, forming a falsifiable local hypothesis, making the smallest grounded edit, and immediately running focused validation.

When should I use Narrow Debug Fix?

Narrow Debug Fix fits situations like: : debugging a failing test; A review comment asking for a narrow fix.

How do I install Narrow Debug Fix in Claude Code?

Run `npx skills add lge-ros2/cloisim --skill narrow-debug-fix -a claude-code`. Or copy the skill folder (.github/skills/narrow-debug-fix in lge-ros2/cloisim) into .claude/skills/narrow-debug-fix in your project. Claude Code loads it when a task matches its description.

How do I install Narrow Debug Fix in Codex?

Run `npx skills add lge-ros2/cloisim --skill narrow-debug-fix -a codex`. Or copy the skill folder (.github/skills/narrow-debug-fix in lge-ros2/cloisim) into .agents/skills/narrow-debug-fix in your project. Codex loads it when a task matches its description.

Can I use Narrow Debug Fix in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add lge-ros2/cloisim --skill narrow-debug-fix -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/narrow-debug-fix, .gemini/skills/narrow-debug-fix, .github/skills/narrow-debug-fix and .opencode/skills/narrow-debug-fix in your project.

What does Narrow Debug Fix need to run?

Going by SKILL.md and its folder, Narrow Debug Fix needs the command-line tools its instructions call (git).

Does Narrow Debug Fix access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Narrow Debug Fix safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Narrow Debug Fix use?

Narrow Debug Fix is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Narrow Debug Fix use?

About 1.2k tokens (SKILL.md is roughly 4.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Narrow Debug Fix?

Skills that share tags, products or a category with Narrow Debug Fix: React Router Bug Fix Workflow (remix-run/react-router, 57k stars), Runtime Debug (vercel/next.js, 143k stars), Root Cause Debugging (jsmastery-pro/skills, 1.4k stars) and Superpowers Systematic Debugging (christopherarter/superpowers-reasonix, 102 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Narrow Debug Fix?

lge-ros2 (a GitHub organization) maintains it in lge-ros2/cloisim, which has 176 GitHub stars. The repository holds 18 skills in this directory. The repository was last updated on October 6, 2026.

Source: lge-ros2/cloisim on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.