AgentSkillsCN

test-quality-verifier

对模糊断言进行测试审计，提升覆盖率，并生成结构化的报告。与@test_quality_verifier配合使用效果更佳。

SKILL.md
--- frontmatter
name: test-quality-verifier
description: Audit tests for vague assertions, improve coverage, and produce a structured report. Pairs well with @test_quality_verifier.
metadata:
  short-description: Verify and improve test quality

Test Quality Verifier

Use this skill when you need to validate that tests are meaningful (not just truthy checks), cover important behaviors, and actually fail when the implementation is wrong.

Quick Use

  • •If multi-agent is enabled, invoke @test_quality_verifier and ask it to audit the current repo's tests and improve them as needed.
  • •If working solo, follow the workflow below.

Workflow

  1. •Identify the test runner and conventions (look for package.json, pyproject.toml, go.mod, Makefile, CI workflows).
  2. •Enumerate tests (rg -n "describe\\(|it\\(|test\\(|def test_|func Test" . is a quick start).
  3. •Find vague assertions:
    • •JS/TS: lone .toBeTruthy(), .toBeDefined() with no value/shape checks; literal==literal asserts.
    • •Python: assert True, assert result with no comparison; pass in tests.
    • •Go: tests that only check err != nil and never validate returned values.
  4. •Run the repo’s existing test/coverage commands if present; otherwise use framework defaults.
  5. •Replace vague asserts with checks on actual values, shapes, and error paths. Add tests for uncovered branches.
  6. •Re-run tests/coverage and ensure the report reflects reality.

Output

Produce a report with:

  • •Files scanned, vague assertions found/fixed, tests added
  • •Coverage numbers if available (and where they came from)
  • •PASS/FAIL with concrete reasons