Debug & triage
5,200 rows from 1,554 repositories
Something is wrong and you do not yet know what.
What Debug & triage skills agree on
1,113 skill files read, by 1,020 of the 1,639 authors on this shelf whose files we hold, 2026-09-06
The middle one of the 136 measured here is ~1.7k tokens long, counted with cl100k_base
Counted by distinct author, so one author publishing three of these counts once. Where a claim sits in fewer files than authors, that is said: a claim held by forty authors across three files is one file people copied, not forty people who agreed. Near-identical wordings are grouped and the other wordings are shown, so the grouping is yours to check.
What they tell the agent to do
- Find root cause before attempting any fix134 of 1020 in 118 filesalso worded as investigate root cause before proposing any fix; find root cause before attempting any fixes
- Create a failing test case before implementing a fix109 of 1020 in 95 filesalso worded as create a failing test case before fixing; write a failing test before fixing
- Read error messages and stack traces completely102 of 1020 in 88 filesalso worded as read error messages and stack traces carefully; Read all error messages and stack traces completely
- Reproduce the issue consistently before investigating90 of 1020 in 77 filesalso worded as reproduce the issue reliably before investigating; reproduce the issue consistently before guessing
- Make the smallest possible change to test a hypothesis90 of 1020 in 76 filesalso worded as make the smallest possible change to test hypothesis; make the smallest possible change to test hypotheses
- Trace data flow backward to find the source84 of 1020 in 70 filesalso worded as trace data flow to find the source of bad values; Trace data flow backward to the source
- Form a single hypothesis before testing78 of 1020 in 64 filesalso worded as form a single specific hypothesis; form a single specific hypothesis before testing
- Implement only one fix at a time76 of 1020 in 63 filesalso worded as implement a single fix at a time; implement one fix at a time
- Question the architecture if three fixes fail73 of 1020 in 59 filesalso worded as question the architecture after three failed fix attempts; stop and question architecture after three failed fixes
- Add diagnostic instrumentation at component boundaries68 of 1020 in 56 filesalso worded as add diagnostic instrumentation at each component boundary; add diagnostic logging at component boundaries
- Compare broken code against working examples68 of 1020 in 57 filesalso worded as find working examples to compare against broken code; identify differences between working and broken code
- Write a regression test before applying the fix62 of 1020 in 55 filesalso worded as verify the fix with a regression test; Write a regression test before fixing
- Check recent changes and git diffs54 of 1020 in 51 filesalso worded as check recent changes using git log; check recent git changes
- Generate three to five ranked falsifiable hypotheses44 of 1020 in 39 filesalso worded as generate three to five falsifiable hypotheses; generate three to five ranked hypotheses before testing
- Verify the fix with the full test suite43 of 1020 in 41 filesalso worded as run the full test suite after fixing; Verify the fix with tests
What they tell it not to do
- Do not propose fixes without root cause investigation116 of 1020 in 102 filesalso worded as do not propose a fix without root cause investigation; Do not apply fixes without root cause investigation
- Do not add multiple fixes at once107 of 1020 in 95 filesalso worded as do not change multiple things at once; do not change multiple variables at once
- Do not bundle refactoring with bug fixes74 of 1020 in 60 filesalso worded as do not bundle refactoring with a fix; do not perform refactoring while fixing a bug
- Do not guess the cause of an issue60 of 1020 in 56 filesalso worded as Do not guess the root cause; do not guess the cause of the issue
- Do not skip error messages or warnings55 of 1020 in 46 filesalso worded as do not skip error messages or stack traces; do not skip reading error messages
- Do not attempt a fourth fix after three failures54 of 1020 in 45 filesalso worded as do not add more fixes after three failed attempts; do not add more fixes after three failures
- Do not fix symptoms instead of root causes52 of 1020 in 48 filesalso worded as do not apply fixes without a confirmed root cause; do not fix symptoms without addressing the root cause
- Do not guess if an issue is not reproducible45 of 1020 in 38 filesalso worded as do not guess if the issue is not reproducible; do not guess at fixes without reproducing the bug
- Do not skip tests34 of 1020 in 28 filesalso worded as do not skip writing a failing test case; do not skip tests or rely on manual verification
- Do not proceed to hypothesize without a feedback loop30 of 1020 in 24 filesalso worded as do not proceed to hypothesize without a loop; do not proceed to hypothesise without a loop
What they expect to be installed
- git263 of 1020 in 251 files
- grep155 of 1020 in 139 files
- gh82 of 1020 in 78 files
- curl79 of 1020
- security62 of 1020 in 48 files
- codesign60 of 1020 in 46 files
- debugger56 of 1020 in 55 files
- env55 of 1020 in 43 files
- jq51 of 1020 in 44 files
- git bisect48 of 1020 in 40 files
What they ask it to produce
- Failing test case121 of 1020 in 104 filesalso worded as failing test; failing test cases
- Root cause analysis105 of 1020 in 101 filesalso worded as root cause hypothesis; root cause report
- Regression test98 of 1020 in 97 filesalso worded as Regression tests; fix with regression test
- Diagnostic logs77 of 1020 in 64 filesalso worded as diagnostic log entry; Diagnostic instrumentation logs
- Documented root cause72 of 1020 in 58 filesalso worded as documented root cause analysis; Root cause documentation
- Ranked hypothesis list40 of 1020 in 34 filesalso worded as ranked list of hypotheses; Hypothesis list
- Diagnostic report36 of 1020also worded as diagnosis report; diagnostic reports
- Triage report35 of 1020 in 32 filesalso worded as triage_report.json; triage summary report
When Debug & triage authors say to reach for one
The situations these authors wrote into their own files, counted out of the same 1,020 authors, with the skills that name each one
- Test failures165 of 1020 in 149 files
- Unexpected behavior143 of 1020 in 127 files
- Build failures99 of 1020 in 84 files
- Performance problems92 of 1020 in 77 files
- Bugs in production83 of 1020 in 68 files
- Integration issues80 of 1020 in 65 files
- User reports a bug60 of 1020 in 54 files
- User says diagnose or debug this30 of 1020 in 24 files
- Diagnosing bugs
- Diagnose
- Skill diagnose
- and 26 more on this shelf
How Debug & triage skills are built
4,551 skill directories by 1,430 authors, read from their repositories’ own file trees 2026-08-05
The middle bundle among those shipping files is 3 files, 15.3 KB beside SKILL.md
Counted by distinct author, same as above, so one author publishing forty template copies counts once. SKILL.md itself is not counted as a file, so a single-file skill is one where that file is the whole skill.
The shape
- SKILL.md is the whole skill573 of 1,430 authors, 2,122 of 4,551 skills
- files ship beside it857 of 1,430 authors, 2,429 of 4,551 skills
- executable scripts ship inside346 of 1,430 authors, 755 of 4,551 skills
The folders they converge on
- references/488 of 1,430 authors, 1,284 of 4,551 skills
- scripts/262 of 1,430 authors, 575 of 4,551 skills
- agents/158 of 1,430 authors, 278 of 4,551 skills
- assets/78 of 1,430 authors, 115 of 4,551 skills
- evals/77 of 1,430 authors, 175 of 4,551 skills
- examples/62 of 1,430 authors, 89 of 4,551 skills
getsentry/sentry-mcp/io.github.getsentry/sentry-mcp MCP server
in 2 stacksno license808★ repoMCP server for Sentry - error monitoring, issue tracking, and debugging for AI assistants
kubesphere/kubesphere/skills/frontend-forge-fe-operations Skill
no license17,025★ repoThe container platform tailored for Kubernetes multi-cloud, datacenter, and edge management ⎈ 🖥 ☁️
WenyuChiou/awesome-agentic-ai-zh/examples/stage-5/tool-calling-tutor Skill
5,144★ repoA trilingual (繁中 / English / 简中) learning roadmap for agentic AI: from LLM basics to multi-agent systems, with 240+ curated resources and hands-on examples. 中文 AI agent 學習地圖。
inkeep/open-knowledge/packages/server/assets/skills/bug-report Skill
3,424★ repoBeautiful, AI-native markdown IDE and LLM wiki
MCP Server for WinDbg Crash Analysis
svnscha/mcp-windbg/io.github.svnscha/mcp-windbg MCP server
1,521★ repoA Model Context Protocol server for Windows crash dump analysis using WinDbg/CDB
noobnooc/agent/skills/better-codex Skill
no license1,374★ repoMy profile & the agent skills I created
Optim-Agent/optim-plans/skills/analyze-and-plan Skill
494★ repoHuman-in-the-loop planning plugin for Claude and Codex: turn ideas into reviewed Markdown plans, record decisions, enforce explicit execution gates, and provide tested controller primitives for safer agent workflows.
VisionForge-OU/foreman/src/foreman/skills/foreman-debug Skill
no license484★ repoA Boris-style agentic orchestrator TUI that supervises headless Claude Code agents through a gated software-delivery pipeline — pointed at any repository.
leiting-eric/DailyBrief/.claude/skills/daily-brief Skill
320★ repoAI 每日新闻简报 · GitHub 热门 + X 热门文章 + 行情技术分析 · 23 个数据源聚合 + LLM 中文摘要 · 本地或 GitHub Actions 部署
qdrant/skills/meta/qdrant-advisor Skill
220★ repoAgent skills for Qdrant vector search: scaling, performance optimization, search quality, monitoring, deployment, model migration, version upgrades, and SDK usage across Python, TypeScript, Rust, Go, .NET, Java
techygarg/lattice/dev-skills/skill-align Skill
176★ repoInstall engineering discipline into any AI coding assistant. Composable skills for design, implementation, review, and team standards. Better process, not just better prompts.
Alibabacloud ecs linux troubleshooting
aliyun/alibabacloud-ecs-troubleshoot-skills/skills/alibabacloud-ecs-linux-troubleshooting Skill
152★ repoTroubleshooting skills for Alibaba Cloud ECS
sendaifun/skills/skills/arcium Skill
126★ repoa public marketplace of all solana-related skills for agents to learn from!
bromoket/x64dbg_mcp/io.github.bromoket/x64dbg MCP server
97★ repoMCP server for x64dbg debugger - 23 mega-tools for AI-powered reverse engineering and debugging
mturac/everything-openai-codex/skills/agent-introspection-debugging Skill
88★ repoEOC: open-source operating system for OpenAI Codex workflows with agents, skills, hooks, rules, memory, safety gates, and cross-harness adapters.
harness/harness-skills/skills/ai-operations Skill
83★ repoA collection of structured AI agent skills that enable Claude Code, Cursor, GitHub Copilot, and other AI coding assistants to create, operate, debug, and govern Harness CI/CD workflows through natural language.
BagelHole/DevOps-Security-Agent-Skills/compliance/auditing/audit-logging Skill
46★ repoAgent-ready DevOps, security, infrastructure, and compliance knowledge base with 80+ skills across Kubernetes, Terraform, AWS/Azure/GCP, AI platform operations, container hardening, SOC2/ISO27001, and incident response—plus ready-to-run scripts, templates, and playbooks for SRE, platform, and security teams.
h0x0er/ebpf-skill Skill
42★An eBPF skill for coding agents
aligundogdu/symfony-hexagonal-skill/skills/symfony-api-response Skill
no license32★ repoClaude Code plugin that enforces hexagonal architecture (ports & adapters) in Symfony projects — with 10 auto-triggered skills, 2 review agents, and progressive refactoring support for existing codebases.
OthmanAdi/langsmith-fetch-skill Skill
27★🔍 AI observability skill for Claude Code. Debug LangChain/LangGraph agents by fetching execution traces from LangSmith Studio directly in your terminal.
ptsilivis/autonomousguy/skills/autosar/autosar-bsw Skill
26★ repoAutonomousGuy — AI skill prompts for embedded automotive engineers (AUTOSAR, MISRA, ISO 26262, ECU)
Maxmedawar/tandem/skills/tandem-agentic-engineering Skill
25★ repoTandem is an MCP bridge that lets a chatbot (Claude.ai or ChatGPT) spin up and drive, interactive Claude Code sessions on your own machine. Runs 2 sessions in a lead/worker loop that works autonomously. Also makes claude code respond to the initial Chatbot through the browser so it can respond back.
Borda/AI-Rig/plugins/cc_develop/skills/debug Skill
24★ repoA collection of personal AI coding assistant configurations, specialist agents, and automated workflows optimized for Python and ML open-source development.
easyzoom/aix-skills/skills/8051-mcu-debug Skill
24★ repoReusable Agent Skills for embedded systems, MCU debugging, firmware workflows, and AI automation. （嵌入式、MCU 调试、固件流程与 AI 自动化的可复用 Agent Skills 集合）。
tortastudios/sentry-instrumentation Skill
24★Sentry instrumentation skill for system-behavior tracking
Fabric data agent perf remediate
PatrickGallucci/fabric-skills/skills/fabric-data-agent-perf-remediate Skill
16★ repoCopilot, Claude AI Skills for Microsoft Fabric
eduwxyz/my-awesome-skills/skills/diagnose Skill
no license16★ repoClaude Code plugin marketplace shipping an SDD pipeline (interview-to-spec, diagnose, spec-approach, tdd, verify) with hooks that turn skill rules into mechanical guarantees.
ArunJRK/audit-flow Skill
15★Claude Code skill for interactive system flow tracing across CODE, API, AUTH, DATA, NETWORK layers. SQLite persistence, Mermaid export, git merge driver.
ASP.NET Core Debugger — MCP Server for AI Agents
magna-nz/aspnetcore-debugger-mcp/io.github.magna-nz/aspnetcore-debugger-mcp MCP server
14★ repoMCP server giving AI agents interactive .NET / ASP.NET Core debugging via netcoredbg + DAP.
vanillaflava/tasknotes-skill/tasknotes Skill
12★ repoAgent skill for task management in Obsidian reads and writes your TaskNotes directly as markdown files. Pairs with llm-wiki-claude-skills for an assisted knowledge + task workflow.
wuemaikblume/dsgvo-skills/claude/skills/dsgvo-auth-and-logging Skill
no license11★ repoDSGVO Compliance Skills für Claude AI / Claude Code — DACH-fokussierte Best-Practice-Sammlung für Drittlandtransfer (Art. 44-49) mit US-CLOUD-Act-Hinweis
kageroumado/swiftui-macos-skill/skills/swiftui-macos Skill
11★ repoAgent skill for writing and reviewing SwiftUI code on macOS — runtime-level understanding of observation, concurrency, view identity, and platform integration
Fuenfgeld/pydantic-ai-skills/skills/pydantic-ai-agents Skill
10★ repoProduction-ready Claude Code skills for building AI agents with Pydantic AI. Includes dependency injection, tools, validators, streaming, multi-agent orchestration, and evaluation framework patterns.
svd-ai-lab/sim-plugin-abaqus/src/sim_plugin_abaqus/_skills/abaqus Skill
10★ repoUse Codex and Claude Code with Abaqus via sim-cli: CAE/noGUI authoring, batch jobs, ODB diagnostics, bounded execution.
Stunspot/CanopyOps/canopyops Skill
no license10★ repoEvidence-bounded cannabis cultivation operations Augment for crop plans, diagnostics, calculations, harvest review, and operating records.
geledek/enterprise-ai-transformation-skills/skills/general-idea-diagnostic Skill
9★ repoAgent skills for navigating AI transformations in enterprises.
akovalion/paranoid-qa/ru/skills/bug-report Skill
9★ repoClaude Code skills that turn an AI agent into a meticulous QA engineer. Evidence or it didn't happen.
voidmatcha/e2e-skills/skills/cypress-debugger Skill
9★ repoAI agent testing toolkit for Playwright and Cypress: generate E2E tests from scratch, review existing specs against 24 anti-patterns (P0/P1/P2 silent-always-pass smells), and debug flaky failures from playwright-report/ or cypress/reports/. Agent Skills for Claude Code and Codex.
HewlettPackard/gl-mcp/io.github.HewlettPackard/greenlake-audit-logs-mcp MCP server
9★ repoHPE GreenLake audit-logs MCP Server
clamp-sh/analytics-skills/skills/analytics-diagnostic-method Skill
8★ repoSkills for AI agents to read web analytics like a senior analyst: diagnose traffic, judge channels, read funnels, declare typed events, and read A/B tests. Tool-maps for Amplitude, Clamp, GA4, Mixpanel, and PostHog.
88lin/computer-care/skills/computer-care Skill
8★ repoCross-platform computer diagnosis and repair Skill for Codex, Claude Code, and compatible agents.
ninthwave-io/ninthwave/skills/decision-triage Skill
8★ repoOrchestrate parallel AI coding into reviewable PRs.
pilot617/awesome-claude-code-plugins/plugins/cw-debug/skills/cw-debug Skill
7★ repoA curated marketplace of high-quality Claude Code plugins for developers who ship production software.
r00tify/crashdx/io.github.r00tify/crashdx MCP server
7★ repoDiagnoses Apple .ips crash reports: ranks likely causes with cited evidence, not just a stack trace.
voderment/AI-UIconfig-skill Skill
6★A reusable skill for turning SwiftUI codebases into a governed, previewable, tokenized UI workflow.
elijeangilles/revops-skills/skills/activity-capture-diagnostic Skill
6★ repoAudit your Salesforce RevOps. Then fix it. A Claude Agent Skills pack for Salesforce admins and RevOps practitioners.
AnmarHani/SwarmVault/skills/swarm-debug Skill
5★ repoA shared knowledge vault + full software-engineering workflow for AI coding agents — run Claude Code and Codex in parallel with one memory, one plan.
Toubat/agentic-debug-mode/skills/agentic-debug-mode Skill
5★ repoRuntime-evidence debugging for coding agents: session-scoped probes, bounded logs/query CLI, and an agent skill.
vladzima/video-debug Skill
5★Debug UI bugs from screencasts: an Agent Skill that extracts key frames via ffmpeg scene detection so coding agents can visually inspect what went wrong.
rafaelghif/antigravity-agents/.agents/skills/advanced-debugger Skill
5★ repoEnterprise-Grade Guardrails, Zero-Assumption Execution, and Quality Gates for Autonomous AI Coding Agents.
APareek89/power-coding Skill
5★Mindful-coding OS for any LLM coding agent (Claude Code, Codex, Gemini CLI) — the Five Forces: session handoff, eval loops, vibe-debugging diagrams, FMEA failure scans + a pre-commit secret gate. Plain markdown, no daemons.
tantiope/datadog-mcp-server/io.github.TANTIOPE/datadog-mcp MCP server
5★ repoFull Datadog API access: monitors, logs, metrics, traces, dashboards, and observability tools
ajaydata-vision/nuclear-bug-fix- Skill
no license4★Single shot Bug Fixing Skill for AI Agents
penspanic/webgpu-metal-debug-kit Skill
4★Agent Skill + Claude Code Plugin for debugging Chrome WebGPU on macOS — Chrome DevTools MCP + Xcode Metal tracing
walterkken/doc-driven-engineering-software-skills/ansys-agent Skill
no license4★ repoDoc-driven AI-agent skills for COMSOL, ANSYS, AutoCAD, and SOLIDWORKS automation.
curtischoutw/claude-institution/institution/skills/debug-protocol Skill
4★ repoAfter distilling Fable 5, the output is an indexed CLAUDE.md + on-demand rules + skills, an executable specification designed for weaker model sessions.
zerocracy/zealot/skills/file-bug-report Skill
4★ repoA set of skills for an extremely disciplined AI agent helping maintain a GitHub repository
cosbort/agent-skills/java-spring-best-practices Skill
4★ repoReusable Agent Skills for AI coding agents — Spring Boot, Java, enterprise best practices
Brattlof/skillet/examples/explain-error Skill
4★ repoPackage manager for MCP servers, Agent Skills (SKILL.md), and Claude Code commands, hooks, subagents, and output styles. Installs into Claude Code, Cursor, Codex, Gemini CLI, Copilot, and Cline.
microhenrio/openocd-mcp/io.github.microhenrio/openocd-mcp MCP server
4★ repoDebug microcontrollers via OpenOCD: flash, breakpoints, memory, registers and variables.