This skill provides mandatory rules for handling test failures during AI-assisted testing.
🧪 AI TEST FAILURE RESOLUTION RULES
When running Odoo tests, analyze any failures and apply exactly one of the three rules below — nothing else:
-
RULE 1 — Custom test fails: The custom module code is broken (e.g. upgrade logic did not preserve the expected workflow). → Fix the custom module CODE so the workflow is correct again. Never weaken or skip the test.
-
RULE 2 — Standard Odoo test fails because of our custom code: Our changes altered a workflow or model that a standard test relied on. → Monkey-patch the standard test from within the custom module so it aligns with the new workflow. Never modify standard Odoo files. Never make the test trivially pass by removing assertions.
-
RULE 3 — Standard Odoo test fails for an unrelated reason: → Do nothing. Skip it and move on.
After applying a fix, re-run the tests and repeat until all Rule-1 and Rule-2 failures are resolved. Finish with a summary of every change made and which rule justified it.