systematic-debugging
Skill0 starsChanged 49 days agoMITView on GitHub
A disciplined debugging protocol to use the moment you hit a bug, test failure, unexpected behavior, or job failure — BEFORE proposing any fix. Iron rule = no fix until the root cause investigation is done. The trigger is the instant you think "let me just try this". Auto-fires on "there's an error", "it's not working", "it's failing".
190 other projects use a version of this.
Adopted by
Most recently taken up by
- DemonDamon/AgenticXtheir own version, first seen Nov 9, 2026
- chrysogonus/power-agentstheir own version, first seen Nov 8, 2026
- willy92wins/DayZ-Modding-Knowledge-Packtheir own version, first seen Nov 8, 2026
- roundpilot/superpowers-antigravitytheir own version, first seen Nov 8, 2026
- REPOZY/superpowers-optimizedtheir own version, first seen Nov 8, 2026
- lingyired/status-triotheir own version, first seen Nov 8, 2026
Used with
Projects with it also use chinese-code-review, chinese-commit-conventions and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
Leodorareluctant259Skillchinese-code-review
中文代码审查规范——在保持专业严谨的同时，用符合国内团队文化的方式给出有效反馈
Also in 8 projects that use this
- Leodorareluctant259Skill
chinese-commit-conventions
中文 Git 提交规范 — 适配国内团队的 commit message 规范和 changelog 自动化
Also in 8 projects that use this
obraSkilldiagnosing-superpowers
Use when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.
Also in 20 projects that use this
mjunaidcaSkillbuilding-chatgpt-apps
Guides creation of ChatGPT Apps with interactive widgets using OpenAI Apps SDK and MCP servers. Use when building ChatGPT custom apps with visual UI components, embedded widgets, or rich interactive experiences. Covers widget architecture, MCP server setup with FastMCP, response metadata, and Developer Mode configuration. NOT when building standard MCP servers without widgets (use building-mcp-servers skill instead).
Also in 4 projects that use this
- mjunaidcaSkill
building-chat-interfaces
Build AI chat interfaces with custom backends, authentication, and context injection. Use when integrating chat UI with AI agents, adding auth to chat, injecting user/page context, or implementing httpOnly cookie proxies. Covers ChatKitServer, useChatKit, and MCP auth patterns. NOT when building simple chatbots without persistence or custom agent integration.
Also in 4 projects that use this
- +2 more from Leodorareluctant259
- +5 more from mjunaidca
Other versions190 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 190
- obra/superpowers★ 314ksame name
- NousResearch/hermes-agent★ 262ksame name
- open-metadata/OpenMetadata★ 16ksame name
- jnMetaCode/superpowers-zh★ 8.4ksame name
- vudovn/ag-kit★ 8.2ksame name
- vudovn/antigravity-kit★ 8.2ksame name
- openai/plugins★ 7.8ksame name
- ChrisWiles/claude-code-showcase★ 6.1ksame name
- farm-fe/farm★ 5.6ksame name
- microsoft/FluidFramework★ 4.9ksame name
180 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- analysis-sweepSkill
- cloud-run-job-deploySkill
- grillingSkill
- pre-deploySkill
- self-auditSkill
- tddSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 177 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Use this discipline *most* under pressure (panic invites guessing.
- Distinctive: line 20
If there is a reference implementation, read all of it.
- Distinctive: line 24
Verify with the smallest change.
- Distinctive: line 26
If you do not know, say "I don't know".
- Distinctive: line 46
If it truly is environmental or timing-related: record the investigation, add proper handling (retry/timeout), and add monitoring/logging …
“Common” is measured: kept by at least 6 in 10 of the 177 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Systematic Debugging (Root Cause Before Fix)
Iron rule: do not fix without investigating the root cause. Guess-fixes aimed at symptoms melt time and spawn new bugs. If Phase 1 is not done, you have not earned the right to name a fix.
Use this discipline most under pressure (panic invites guessing; systematic is faster in the end: measured 15-30 min vs. 2-3 hours of flailing).
Phase 1: Root Cause Investigation (no fix proposals yet)
- Read the error message to the end (full stack trace, line numbers, error code — the answer is often written there).
- Reproduce reliably (if you cannot reproduce, the next move is data collection, not a fix).
- Suspect the most recent change (git diff, recent commits, dependencies, config, environment differences).
- In multi-layer systems, instrument the boundaries before judging (log the input/output of each component, run once, get evidence for which layer breaks, then dig only there). A gap in the logs is itself evidence.
- Trace bad values back to their source (fix at the origin, not where the symptom surfaces).
Phase 2: Pattern Analysis
- Find a working similar case in the same codebase and enumerate every difference from the broken one ("that can't be relevant" is banned).
- If there is a reference implementation, read all of it. Skimming to "borrow the pattern" is how bugs are born.
Phase 3: Hypothesis and Minimal Test
- State each hypothesis in one sentence ("I think X is the cause, because Y").
- Verify with the smallest change. Do not fix several things at once (you lose track of what worked).
- Wrong guess -> new hypothesis. Do not stack fixes on top of each other.
- If you do not know, say "I don't know". Do not fix while bluffing.
Phase 4: Implementation
- Write the failing reproduction test first (a fix without a test does not stick).
- Make one fix, aimed at the root cause (no "while I'm here" refactors mixed in).
- Turn the test green and confirm no other tests broke.
- If a fix fails three times, stop. That is not a wrong guess, it is an architecture problem. Consult a human on the design before attempt #4 (this is your human checkpoint — see the Loop Engineering doc).
Stop signs (think any one of these -> go back to Phase 1)
"Just fix it and investigate later", "let me change X and see", "change several things and test together", "probably X so let me just fix it", "not fully sure but this might work", "just one more fix attempt (already failed twice)".
Excuses vs. reality
| Excuse | Reality |
| It's a simple bug, no protocol needed | Simple bugs have root causes too. If it's simple, the protocol is fast |
| It's urgent, skip the protocol | Systematic beats trial-and-error on wall-clock time |
| I'll investigate after it's fixed | The first move sets the shape. Do it right from the start |
| I'll write the test later | A fix without a test recurs |
If the investigation concludes "no root cause"
If it truly is environmental or timing-related: record the investigation, add proper handling (retry/timeout), and add monitoring/logging for the future. But 95% of "no root cause" is insufficient investigation.
<!-- Derived from the systematic-debugging skill in obra/superpowers (MIT, Copyright (c) 2025 Jesse Vincent), condensed and adjusted. See THIRD_PARTY_NOTICES.md. -->Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.