systematic-debugging: a Claude Code skill by Jesse Vincent
For developers whose AI keeps guessing at fixes and who want the root cause found first. Skip it if you just need a typo or one-line slip corrected.
- Score
- 84of 100
- Rank
- #31of 239
- Installs
- 33Kon skills.sh
- Security
- Mixed 5 scanners
- Updated
- 2026
- Context
- ~27tokens always
Install
npx skills add https://github.com/obra/superpowers --skill systematic-debugging Run it in your terminal. View the source
Using the Claude app rather than Claude Code? You can add a skill there too. How to add it in the app
What does systematic-debugging do?
systematic-debugging is a Claude skill that forbids proposing a fix before the root cause is found. It runs four phases in order. First, read the error and stack trace fully, reproduce the bug, check recent diffs and commits, and log what enters and leaves each component to see where data goes wrong. Second, find similar working code and list every difference. Third, state one hypothesis and test it with the smallest change. Fourth, write a failing test, make one fix and verify it. If three fixes fail, Claude must stop and discuss the architecture with you. It refers to helper files on backward tracing and condition-based waiting.
When should you use systematic-debugging?
Use it when
- A test fails and the cause is unclear
- Several quick fixes have already failed
- A build or CI pipeline breaks across many steps
Skip it when
- The fix is obvious and tiny
- You want Claude to patch first and explain later
What it needs and what it can touch
What it can do
- Runs shell commands and tests
- Reads and edits project files
From reading its instructions, not from running it.
Why it scores 84
Strongest on upkeep (100 of 100), weakest on trust (62).
Is systematic-debugging safe to install?
At least one scanner raised a medium warning. Read the results below. Scanners often miss what other scanners catch, and a skill can change after a scan. Read its source before you install it.
| Scanner | Result | Date |
|---|---|---|
| Gen Agent Trust Hub | Medium risk | Nov 4, 2026 |
| Socket | Passed | Nov 5, 2026 |
| Snyk | Low risk | Nov 4, 2026 |
| Runlayer | Low risk | Feb 26, 2026 |
| ZeroLeaks | Passed | Apr 16, 2026 |
Details
- Installs
- 329,528
- Repository stars
- 316,405
- Last repository change
- Tokens always loaded
- about 27
- Tokens when used
- about 2,341
- License
- MIT
- Data confidence
- High
Did it do the job for you?
Alternatives to systematic-debugging
All skills for this job-
Get tddtdd
For developers who want Claude to build features test first, one failing test at a time.
- Get writing-plans
writing-plans
For developers who give Claude a spec and want a test-first build plan before any code is written.
-
Get vue-debug-guidesvue-debug-guides
For Vue 3 developers chasing runtime errors, warnings, reactivity bugs or hydration mismatches.
-
Get PUAPUA
For developers tired of agents that give up or claim fixes without proof.
systematic-debugging