AGENT SKILL

ws-tdd-review

Audit existing test code for quality, depth, and adherence to TDD principles.

License 许可
MIT
所属仓库
vempio/WhittleSpec
最近更新
2026-10-06
Stars
★ 4
Forks
0

01 前言 Front Matter

name
ws-tdd-review
description
Audit existing test code for quality, depth, and adherence to TDD principles.

03 SKILL.md 正文

ws.tdd.review

Load shared context: ws.tdd._meta/SKILL.md — the sibling skill in the same skills directory this skill loaded from (resolves under any clone name, install location, or skills-root override). Also load: ws.tdd._meta/test-quality-checklist.md (step 3's criteria) and ws.tdd._meta/debt-protocol.md (the debt scan that opens it).

Purpose

Review existing test code against TDD quality standards. Distinct from /ws.review (reviews specifications for testability) -- this reviews actual test code for quality.

Process

1. Scope

Ask: "Which test files or modules should I review?" (or confirm if specified).

2. Depth

Default: thorough. Examine every test file in scope, read each test function. Most common complaint: superficial review missing issues careful reading would catch.

If scope large (> 20 files), state and propose batching: "N files in scope. I'll review in batches of 8, starting with [area]."

3. Evaluate

Work through ws.tdd._meta/test-quality-checklist.md, against ws.tdd._meta/SKILL.md § Quality Standards. The criteria and the adversarial pass both live in that file, so this skill and ws.fix hold one bar rather than two paraphrases of it.

The adversarial pass is not optional. A checklist run without it reports what the tests say; it does not establish that they would catch a wrong implementation.

4. Report

## Test Quality Review: [scope]

### Summary

- **Files reviewed**: [count]
- **Tests reviewed**: [count]
- **Critical issues**: [count] (success theater, broken assertions)
- **Improvements**: [count] (depth, structure, documentation)

### Critical Issues

[Tests not actually validating anything, including untested state clearing paths]

- [File:line] `test_name`: [what's wrong, why it matters]

### Depth Issues

[Assertions checking too shallowly]

- [File:line] `test_name`: checks [what it checks], should check [what it should check]

### Structural Issues

[AAA violations, lumped behaviors, poor naming]

### Dead Code & Usage

- [File:line] `function_name`: defined and tested but never called from production flow

### Test Level

- [File:line] `test_name`: unit-tests coordination logic, should be integration-tested (or vice versa)

### BDD Quality (when .feature files exist)

- [File:line] `scenario_name`: [what's wrong]

### Task Completion Audit (when task file and .feature files exist)

| Scenario | Task | Task Status | Scenario Status | Finding |
|----------|------|-------------|-----------------|---------|
| Generate pending certs | 2b | [x] | active | OK |
| Field validation x3 | 2a | [x] | @todo | CRITICAL: task claims done |
| Send with attachment | 3a | [~] | active | OK (in progress) |
| Full pipeline | -- | -- | @todo | Orphan: no task claims this |

- **Critical**: [count] scenarios still @todo under completed tasks
- **Orphans**: [count] scenarios not traced to any task
- **Pending**: [count] scenarios awaiting activation for incomplete tasks

### BDD-Documentation Correspondence (when operator docs and .feature files exist)

- **Documented workflows without BDD**: [list workflow steps with no matching scenario]
- **BDD scenarios without documentation**: [list scenarios exercising undocumented workflows]
- **CLI usability gaps**: [missing @cli scenarios for --help, unknown subcommand]

### Traceability (when SDD context available)

- **Criteria with tests**: [list with AC identifiers]
- **Criteria without tests**: [list -- these need `/ws.tdd.idea`]
- **Tests without AC reference**: [list -- typically fine; flag only if scope creep]

### Recommendations

[Prioritized list, most impactful first]

Output

Report following template above, delivered for discussion. Report is conversation starter, not final verdict -- human may disagree with severity or priorities.

Relationship to /ws.review

Complementary: /ws.review reviews specifications for testability. /ws.tdd.review reviews actual test code for quality. Both invoked by /ws.fix during quality passes.

Anti-Patterns

  • Reviewing spec testability (that's /ws.review)
  • Suggesting new tests to write (that's /ws.tdd.idea)
  • Rewriting tests during review (review first, fix separately)

04 同仓库更多 Skills