grade-tests

What it does

Grades a specified set of test methods individually and produces a concise table mapping each test (fully-qualified name) to a letter grade (A–F), a score band, and a one-line note — designed to be posted as a PR comment. The skill does not discover tests on its own; the caller provides the test methods to grade.

How it works

1. Validate input to ensure an explicit list of test methods or file paths with a grade instruction is provided. If ambiguous, respond with a request for explicit scope.
2. Detect language and load language-specific extension via the test-analysis-extensions skill and the matching extension file (e.g., extensions/dotnet.md, extensions/python.md).
3. For each listed test, resolve the test body either from provided inline body or by reading the file and locating the fully-qualified method; if not found, mark as N/A — method not found.
4. Score each test starting at A (90–100) and apply deductions based on observable issues in the test body, producing three sub-grades: Assertion strength, Structure & focus, and Anti-pattern hygiene, following the detailed guidance in the skill.
5. Compute final sub-grades using the provided rules, derive an overall grade by converting sub-grades to numeric points (A=4, B=3, C=2, D=1, F=0) with the weighted formula: 0.45 × Assertion + 0.30 × Anti-pattern + 0.25 × Structure, then cap the overall grade to the worst sub-grade. If any sub-grade is F, the overall grade is F, etc.
6. Build the note as a short sentence (≤ 120 characters) explaining the primary reason for the grade, with examples like those in the Step 4 description.
7. Produce a report with two sections: Summary and Per-test table, formatted in Markdown. The table lists Test, Grade, Band, and Notes for each graded test.

When to use it

• When a PR automation workflow needs to post a comment grading the tests introduced or changed in a PR.
• When a reviewer has a specific list of tests and wants per-test verdicts instead of a full suite report.
• When a maintainer wants to triage which tests deserve follow-up improvements.

What it can touch

• Uses language-specific extensions loaded via test-analysis-extensions to determine assertion APIs and idioms for the target language.

Caveats

• The skill does not discover tests on its own; it requires explicit test method inputs or file lists with clear grading scope.
• Production code availability affects whether certain findings can be inferred; if production code is unavailable, mark related findings as "Unverified" in the summary.
• The skill is intended for per-test grading and should not be used for full-suite audits or test-writing tasks.