- Home
- Skills
- Testing & Debugging
- Autonomous Bug Fixer
Autonomous Bug Fixer
Autonomous test-driven debugging agent that iterates on code fixes until your test suite passes.
$3
Works with the AI Agents you already use
Autonomous Bug Fixer
Example session with this skill installed
The test suite is failing on the user authentication module. Fix the code so all tests pass.
- Read your context and instructions
- Compiled the autonomous bug fixer
I have initiated the autonomous debugging loop. After running bash scripts/run-tests.sh, I identified a failure in auth_service.py where a KeyError was triggered during token validation. Based on references/debugging-strategy.md, I hypothesized that the middleware was not handling missing headers.
Debugging Session Log
| Iteration | Action | Result | Hypothesis |
|---|---|---|---|
| 1 | Run scripts/run-tests.sh | Fail (1 error) | Missing 'Authorization' header in request dict. |
| 2 | Edit auth_service.py | Pass | Used .get() with default value to prevent KeyError. |
# auth_service.py refactored snippet
def validate_token(request):
# Fix: Replaced direct access with .get to handle missing headers
auth_header = request.headers.get('Authorization', None)
if not auth_header:
return False
return verify_jwt(auth_header)
All tests are passing. Nothing to fix.
Next steps
- Review the change in
auth_service.pyto ensure it matches your architectural patterns. - Commit the fix to your local branch.
- Run the full integration suite to ensure no regressions in downstream services.
Connects securely to your tools. The creator never sees your data.
About this skill
Manual debugging is a time sink that pulls developers away from building features. This skill automates the cycle of running tests, analyzing failures, and applying code fixes until the suite passes.
What it does
- Autonomous testing executes your local test scripts to find failures.
- Failure analysis interprets test output and consults local debugging guides to find the root cause.
- Hypothesis generation explains the likely bug before any code is modified.
- Self-correcting edits applies minimal code changes and verifies them by rerunning the tests.
- Safety boundaries prevents infinite loops and ensures test cases are never deleted to force a pass.
How it works
- Initialize by running the project test script.
- Analyze failures using the provided debugging strategy reference.
- Iterate through code edits and re-testing for up to five cycles.
- Finalize once the suite passes or provide a detailed report if the bug persists.
Frameworks & tools
This skill works with any stack that uses a bash scripts/run-tests.sh entry point. It relies on a local references/debugging-strategy.md file to guide its logic.
Why this beats prompting it yourself
Generic LLMs often hallucinate fixes or rewrite entire files unnecessarily. This skill is constrained to minimal edits and follows a strict test-driven loop that verifies every change before concluding.
Use cases
- Fix breaking changes after a dependency update or refactor.
- Resolve failing unit tests in a CI-like local loop.
- Debug complex logic errors by following a predefined strategy.
- Clean up edge cases in a newly developed feature.
Known limitations
Requires a scripts/run-tests.sh file and a references/debugging-strategy.md file. Limited to five iteration attempts per session.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.
Requires bash scripts/run-tests.sh and references/debugging-strategy.md to exist. Hard limit of 5 attempts. Cannot fix issues that require deleting tests.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago