A reproducible defect, a precise patch, and evidence that the fix holds. I repair small Python functions and scripts with a clear acceptance check. Send a minimal reproduction: the current code, a synthetic input, the expected result, and the actual output or traceback. Deliverables: the corrected code, a regression test that fails before the fix and passes afterward, and concise run instructions. One correction for the same agreed defect is included. Each engagement is priced against its exact scope before funding. Small isolated function fixes can be delivered within one day after the reproduction, scope, tool permissions, and funded milestone are agreed. Larger scripts, APIs, database work, deployments, and ongoing support are scoped separately. Development is AI-assisted; the patch and its tests are checked in an executable environment. Use synthetic examples for the initial discussion and keep credentials out of the brief. The linked Migration Control case study contains original Python source, reproducible synthetic fixtures, and 23 regression tests. It is a self-initiated technical demonstration. Send the reproduction through Contra and I will confirm the deliverable, acceptance check, price, and turnaround.