code-testing-quality
Skill0 starsChanged 9 months agoMITView on GitHub
Unit testing principles, test structure patterns, test doubles strategy, naming conventions, testing antipatterns, and coverage guidance. Use when writing tests, reviewing test quality, evaluating test coverage, choosing test doubles, fixing flaky tests, or assessing test structure during cook and serve phases. Enhances taster agent during TDD cycle.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- code-antipatternsSkill
- code-namingSkill
- code-plan-auditSkill
- code-pruningSkill
- code-quality-foundationsSkill
- code-readabilitySkill
- code-reviewSkill
- code-scope-boundariesSkill
- code-yagniSkill
- error-handling-patternsSkill
- refactoring-patternsSkill
- software-tradeoffsSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 50
The failure message test: When a test fails, can you understand the bug from the test name alone?
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Code Testing Quality
Testing Quality Checklist
- Each test verifies one behavior (single assertion theme)
- Tests are independent — no shared mutable state, runs in any order
- Test names describe the scenario and expected outcome
- Arrange/Act/Assert sections are visually distinct
- No logic in tests (no conditionals, loops, or try-catch)
- Test doubles used only at architectural boundaries
- Flaky indicators absent (sleeps, real clocks, network calls, file I/O)
- Edge cases covered (nulls, empty collections, boundaries, error paths)
- Tests verify behavior, not implementation details
- Refactoring production code doesn't break tests (unless behavior changes)
Three Properties of Good Tests
| Property | Meaning | Violation Signal |
| Isolated | No test affects another; runs alone or in any order | Tests pass individually but fail together |
| Repeatable | Same result every time, on any machine | "Works on my machine," intermittent failures |
| Clear | Failure message pinpoints the problem | Must read test code to understand what broke |
Test Doubles Decision Table
| Situation | Recommended Double | Why |
| External API or network call | Stub | Isolation from external non-determinism |
| Database in integration test | Fake (in-memory DB) | Realistic behavior, fast execution |
| Verifying a notification was sent | Mock | The side effect is the behavior |
| Complex collaborator with state | Fake | Maintains realistic internal consistency |
| Legacy code you can't refactor yet | Spy | Observe without modifying production code |
| Internal collaborator (same module) | Real object | Prefer real dependencies for internal code |
Rule of thumb: Mock at architectural boundaries (network, database, filesystem, external services). Use real objects for internal collaborators.
Test Naming Patterns
| Pattern | Example | Best For |
method_scenario_expectedResult | withdraw_insufficientFunds_throwsException | Unit tests with clear method focus |
should [outcome] when [condition] | should reject withdrawal when funds insufficient | Behavior-focused tests |
given_when_then | givenInsufficientFunds_whenWithdraw_thenThrows | BDD-style tests |
The failure message test: When a test fails, can you understand the bug from the test name alone? If yes, the name is good.
Testing Antipatterns
| Antipattern | Symptom | Severity | Fix |
| Flaky tests | Pass/fail without code changes | Critical | Remove non-determinism: no sleeps, no real clocks, no shared state |
| Test interdependence | Tests pass alone, fail together | Critical | Each test creates/destroys its own data |
| Testing implementation | Refactoring breaks tests | Major | Assert on outputs and side effects, not internal method calls |
| Logic in tests | Conditionals, loops, try-catch in test code | Major | Tests should be straight-line code |
| Slow tests | Suite takes minutes, developers skip it | Major | Isolate from I/O; use fakes; parallelize |
| Obscure tests | Must read full test to understand intent | Moderate | Descriptive names, visible AAA structure |
| Fragile assertions | Tests break on irrelevant output changes | Moderate | Assert on relevant fields only |
| Liar tests | Tests that always pass regardless of behavior | Critical | Verify tests fail when behavior is wrong (mutation testing) |
| Giant arrange | 50 lines of setup, 1 line of act | Moderate | Use builders/factories; consider whether unit is too large |
| Snapshot abuse | Blindly approving snapshot updates | Moderate | Prefer explicit assertions for logic |
Coverage Guidance
| Coverage Target | Context | Rationale |
| >80% of critical paths | Most projects | Balances thoroughness with diminishing returns |
| >90% | Safety-critical, financial | Higher stakes justify investment |
| 60-70% | Early-stage, prototypes | Focus on core behavior first |
Beyond line coverage: Branch coverage is more meaningful. Mutation testing reveals "liar tests" — tests that run code but don't verify it. A high mutation score (>80%) indicates tests that actually catch defects.
"Why Is This Test Bad?"
| Symptom | Likely Antipattern | Action |
| Test breaks after refactoring (behavior unchanged) | Testing implementation | Assert on outputs, not method calls |
| Test fails intermittently | Flaky test | Remove non-determinism (time, state, I/O) |
| Can't understand failure from test name | Obscure test | Rename to describe scenario + expected outcome |
| 50+ lines of setup | Giant arrange | Use builders; consider splitting the unit |
| Tests pass but bugs ship | Liar test / weak assertions | Add mutation testing |
| Test suite takes >5 minutes | Slow tests | Replace I/O with fakes; parallelize |
Checklists
Before Writing a Test
- Identify the specific behavior to verify (not the implementation)
- Determine test level: unit, integration, or E2E
- Choose test double strategy (prefer real objects; mock at boundaries)
- Plan for edge cases and error paths
Test Quality Self-Review
- Each test has one clear reason to fail
- Test names describe scenario and expected outcome
- No shared mutable state between tests
- No non-determinism (time, randomness, network, filesystem)
- Assertions verify behavior, not implementation
- Test fails when the behavior it guards is broken
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.