Prompt file imported from Harsha-Vardhan-69/Resuable-Automation-Selenium (
.github/prompts/post-failure-fix-loop.prompt.md). Fill in{{input}}before use. Copyright stays with the author.
Related skills: artifact-triage, targeted-test-verification, locator-wait-stabilization, verification-impact-mapper.
Use .github/workflow/Framework.md as the source of truth and move from evidence to the smallest verified fix without widening scope unnecessarily.
Requirements:
- Start from the most relevant failure evidence first, such as surefire output, report artifacts, screenshots, logs, or the specific failing test named by the user.
- Diagnose the likely root cause before editing any code.
- Keep the fix limited to the direct failure path unless shared framework risk clearly requires a broader change.
- Re-run the narrowest useful Maven or TestNG verification first, then expand only when the changed area justifies it.
- If the issue is environment-only, third-party-site drift, or missing evidence rather than a clear code defect, state that explicitly instead of forcing a code change.
- Summarize root cause, changed files, verification run, and any remaining risk.
Failure loop context: {{input}}