Test Adequacy Guard
Version updated for https://github.com/Ostico/test-guard to version v1.4.2.
- This action is used across all versions by 1 repositories.
Action Type
This is a Composite action.
Go to the GitHub Marketplace to find the latest changes.
Action Summary
Test-Guard is a GitHub Action designed to evaluate whether a pull request (PR) includes adequate test coverage, going beyond simple coverage percentages. It combines diff coverage analysis, heuristic test-file matching across multiple programming languages, and AI-powered per-file evaluation to provide a definitive pass/fail assessment for each modified file in a PR. By automating the detection of untested changes and offering precise insights, it streamlines code review processes and ensures robust testing practices.
What’s Changed
🧪 Test-Guard
A GitHub Action that gates pull requests on test adequacy. Three layers — coverage analysis, heuristic file matching across 19 languages, and AI-powered per-file evaluation — so untested code stops slipping through review.
Quick Start
- uses: ostico/test-guard@v1
with:
coverage-file: coverage.xml
No external API keys needed — uses the GitHub Models API with your existing GITHUB_TOKEN.
What’s Fixed in v1.4.2
🔧 Token Budget Constants (Bug 1)
GitHub Models free-tier enforces an 8,000-token hard limit, but the previous constants allowed batches up to ~7,991 tokens — right at the edge, causing sporadic 413 Request body too large errors in CI.
Fixed constants:
| Constant | Before | After |
|---|---|---|
_CHARS_PER_TOKEN | 4 | 3 |
_INPUT_TOKEN_LIMIT | 8,192 | 8,000 |
_SYSTEM_OVERHEAD_TOKENS | 700 | 800 |
_SAFETY_FACTOR | 0.80 | 0.85 |
New budget: 6,120 tokens / 18,360 chars per batch — well within the 8,000-token limit.
🔄 413 Model Fallback (Bug 2)
When a batch still exceeds the token limit after truncation retry, the model loop now falls back to the next model (e.g. gpt-4.1-mini → gpt-4.1-nano) instead of stopping with a SKIP verdict. Previously only 403 Forbidden triggered model escalation; now 413 does too.
Changelog
- 🐛
fix: correct token budget constants for GitHub Models 8,000-token limit - 🐛
fix: escalate to fallback model on 413 after truncation retry fails
Full Changelog: https://github.com/Ostico/test-guard/compare/v1.4.1...v1.4.2