Add Copilot Autofix Workflow for Test Failure Remediation

[muldercw]

Summary

This PR adds a Copilot Autofix GitHub Actions workflow that automatically analyzes test failures on pull requests and posts fix suggestions directly on the failing PR.

What Changed (Latest)

The workflow was completely reworked from a generic "scan code and create issues" approach to a targeted test-failure autofix model:

Before	After
Triggered on every push to master	Triggers only on Run tests workflow completion
Created new GitHub issues for detected problems	Posts fix suggestions on the existing PR
Generic code analysis (bugs, smells, tech debt)	Focused on making failing tests pass
Named copilot-auto-issues	Renamed to copilot-autofix

Workflow Behavior

Triggers:

• workflow_run completion for the Run tests workflow (on master branch)
• Manual dispatch via Actions tab

On test failure, Copilot will:

1. Gate on failure - if tests passed, exit immediately (no noise)
2. Find the associated PR - only acts on PRs, skips direct pushes
3. Download test artifacts - parses JUnit XML (test-results-*) for failing test paths, errors, stack traces, and platform info
4. Analyze root causes - checks out the failing commit, traces errors back to source code
5. Post fix suggestions on the PR - using GitHub suggestion blocks (one-click accept) with:
   • Link to the failed workflow run
   • Test failure count and affected platforms
   • Root cause explanation for each failure group
   • Exact code suggestions with suggestion blocks
6. Stay scoped - only fixes what's needed to make tests pass; no unrelated refactoring

Scope rules:

• Fix only what's needed to make failing tests pass
• Never weaken test assertions - fix the source instead
• Group multiple failures with the same root cause
• Flag unclear or risky fixes rather than guessing
• Note merge conflicts for manual resolution

Files

File	Purpose
.github/workflows/copilot-autofix.md	Workflow definition (human-editable source)
.github/workflows/copilot-autofix.lock.yml	Compiled GitHub Actions workflow (auto-generated by gh aw compile)
.github/aw/actions-lock.json	Pinned action versions for reproducible builds
.gitattributes	Marks .lock.yml as linguist-generated, merge strategy ours

Security

• Fork validation: workflow_run triggers guarded against fork-based attacks
• Zizmor annotations: Dangerous trigger usage documented
• Minimal permissions: contents: read, pull-requests: read, actions: read
• Rate limiting: Max 10 comments per workflow run
• Input sanitization: All user-facing text validated and sanitized

Testing

• Workflow compiles successfully via gh aw compile
• Can be tested manually via gh aw run copilot-autofix
• Will activate automatically on the next Run tests failure on a PR

Related

• Uses gh-aw v0.50.4
• Pinned actions: actions/github-script@v8, github/gh-aw/actions/setup@v0.50.4

[github-actions]

Package	Line Rate	Health
clarifai	45%	❌
clarifai.cli	69%	➖
clarifai.cli.templates	46%	❌
clarifai.client	65%	➖
clarifai.client.auth	67%	➖
clarifai.constants	100%	✔
clarifai.datasets	100%	✔
clarifai.datasets.export	80%	✔
clarifai.datasets.upload	75%	➖
clarifai.datasets.upload.loaders	37%	❌
clarifai.models	100%	✔
clarifai.rag	0%	❌
clarifai.runners	52%	➖
clarifai.runners.models	64%	➖
clarifai.runners.pipeline_steps	39%	❌
clarifai.runners.pipelines	71%	➖
clarifai.runners.utils	62%	➖
clarifai.runners.utils.data_types	72%	➖
clarifai.schema	100%	✔
clarifai.urls	58%	➖
clarifai.utils	61%	➖
clarifai.utils.evaluation	16%	❌
clarifai.workflows	95%	✔
Summary	62% (10128 / 16415)	➖

Minimum allowed line rate is 50%