- Home
- Skills
- Agents & Orchestration
- Pytest Bug Fix Verifier — Reproduce, Fix & Prove
Works with the AI Agents you already use
Pytest Bug Fix Verifier — Reproduce, Fix & Prove
Make an AI coding agent reproduce a failing pytest, apply a minimal fix, rerun target and regression tests, and leave a reviewable evidence pack.
Free
Pytest Bug Fix Verifier — Reproduce, Fix & Prove
Example session with this skill installed
Fix the ZeroDivisionError in tests/test_math.py::test_divide_by_zero and verify it with a regression run of the full math suite.
- Read your context and instructions
- Compiled the pytest bug fix
The original failing pytest node was reproduced with exit code 1, a minimal fix was applied, and the same target was revalidated successfully. The required regression run also passed. Final status: PASS. The Pytest Evidence Pack records the execution evidence, changed files, runtime context, status_reason_code: VERIFIED, and a deterministic next_action.
Connects securely to your tools. The creator never sees your data.
About this skill
AI coding agents can claim a Python bug is fixed after changing code without reproducing the original failure or running the relevant regression tests. This Skill enforces a pytest-specific workflow: observe the failure, identify the cause, keep the change narrow, rerun the target and related tests, and report PASS, FAIL or NEED_USER with execution evidence.
The problem
Developers often struggle to verify that a bug fix actually works without manual trial and error. Relying on "it looks right" or unchecked test runs leads to regressions and unverified patches in production.
What it does
- Reproduces specific pytest failures and records verifiable pre-fix execution data.
- Enforces a minimal-fix policy by tracking exactly which files changed and why.
- Verifies repairs by rerunning the original failing tests under identical environment constraints.
- Runs mandatory regression tests to ensure the fix did not break existing functionality.
- Generates a Pytest Evidence Pack JSON containing execution receipts, diff summaries, and root cause analysis.
Frameworks & tools
Python, pytest, and a specialized pytest_evidence.py collector script.
Why this beats prompting it yourself
Generic prompts often hallucinate test passes or skip the critical step of reproducing the failure first. This skill uses a deterministic collector to prove execution, preventing the agent from claiming a fix is successful without real process exit codes and node-ID verification.
Use cases
- Repairing a specific failing test case in a legacy Python codebase.
- Verifying a security patch doesn't break dependent logic via regression testing.
- Creating an audit trail for bug fixes in regulated or high-stakes environments.
- Validating PR contributions by requiring structured evidence of a passing test suite.
Known limitations
Not for greenfield projects or large refactors. Requires an existing, runnable pytest environment and explicit user trust of the project code and plugins.
Requires an existing runnable pytest environment. Not for greenfield projects or broad refactors. Tests that require API keys, tokens, passwords, database credentials, private keys, or other secrets supplied through environment variables are out of scope. Execution is not sandboxed and requires trusted project code.
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 any payment.
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
2 uses
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 2 months ago
- Free with an Agensi account
Requires
Permissions
File Scopes
Uses shell access to run controlled pytest commands, read access to inspect Python code and pytest configuration, and write access only for minimal bug fixes and local Evidence Pack output. The collector does not require browser, network, or environment-variable access. Tests that require secrets or credentials are out of scope for this version.