Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Test-Driven Development (TDD) · dokhacgiakhoa | DecimalAI

dokhacgiakhoa/test-driven-development

SkillScore91Maintained

Use when implementing any feature or bugfix, before writing implementation code

Testingby dokhacgiakhoa

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Coding pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/dokhacgiakhoa-test-driven-development/SKILL.md

decimalai skills pull dokhacgiakhoa-test-driven-development --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/dokhacgiakhoa-test-driven-development/SKILL.md https://app.decimal.ai/s/dokhacgiakhoa-test-driven-development/SKILL.md

Raw file: /s/dokhacgiakhoa-test-driven-development/SKILL.md ↗ · machine-readable summary at /s/dokhacgiakhoa-test-driven-development.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 6d ago

Yes — +50% lift-10% tokens0% turns91% of whole cases pass with skill (20/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-01	✗→✓	▲ Improved	4%	0%
case-02	✗→✓	▲ Improved	-34%	0%
case-03	✗→✓	▲ Improved	-6%	0%
case-04	✗→✓	▲ Improved	-3%	0%
case-06	✗→✓	▲ Improved	-38%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/10/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

dokhacgiakhoa / agent-skills-4-vibe-coding-cli

.agent/skills/test-driven-development/SKILL.md

View on GitHub ↗

License NOASSERTIONImported 8/10/2026Version v1

Skill contents

SKILL.md

Test-Driven Development (TDD)

Overview

Write the test first. Watch it fail. Write minimal code to pass.

Core principle: If you didn't watch the test fail, you don't know if it tests the right thing.

Violating the letter of the rules is violating the spirit of the rules.

When to Use

Always:

New features
Bug fixes
Refactoring
Behavior changes

Exceptions (ask your human partner):

Throwaway prototypes
Generated code
Configuration files

Thinking "skip TDD just this once"? Stop. That's rationalization.

The Iron Law

NO PRODUCTION CODE WITHOUT A FAILING TEST FIRST

Write code before the test? Delete it. Start over.

No exceptions:

Don't keep it as "reference"
Don't "adapt" it while writing tests
Don't look at it
Delete means delete

Implement fresh from tests. Period.

Red-Green-Refactor

dot
digraph tdd_cycle {
    rankdir=LR;
    red [label="RED\nWrite failing test", shape=box, style=filled, fillcolor="#ffcccc"];
    verify_red [label="Verify fails\ncorrectly", shape=diamond];
    green [label="GREEN\nMinimal code", shape=box, style=filled, fillcolor="#ccffcc"];
    verify_green [label="Verify passes\nAll green", shape=diamond];
    refactor [label="REFACTOR\nClean up", shape=box, style=filled, fillcolor="#ccccff"];
    next [label="Next", shape=ellipse];

    red -> verify_red;
    verify_red -> green [label="yes"];
    verify_red -> red [label="wrong\nfailure"];
    green -> verify_green;
    verify_green -> refactor [label="yes"];
    verify_green -> green [label="no"];
    refactor -> verify_green [label="stay\ngreen"];
    verify_green -> next;
    next -> red;
}

🧠 Knowledge Modules (Fractal Skills)

1. RED - Write Failing Test

2. Verify RED - Watch It Fail

3. GREEN - Minimal Code

4. Verify GREEN - Watch It Pass

5. REFACTOR - Clean Up

6. Repeat

Want to see if dokhacgiakhoa/test-driven-development would activate on your phrasing? Try it before you install — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/30/2026, 08:31 UTC · 22 cases

Lift

+50%

91% vs 41%

Agent turns

0%

1.0 vs 1.0

Tokens

-10%

1,703.45 vs 1,899.59

fail → pass · 11pass → pass · 9fail → fail · 2

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I'm starting work on a new user authentication module for our web application. Can you provide a standard step-by-step development guide detailing the complete cycle I should follow from test creation through refactoring?	fail→pass	22,969	18,566	-19%	1	1	0%	3,019	3,148	+4%	0	0	—
case-02 I just realized I wrote a 40-line data filtering function before writing any test cases for it. What is the mandatory protocol for handling code created prior to tests? Please detail the explicit steps I should take with this existing snippet.	fail→pass	20,998	11,076	-47%	1	1	0%	2,580	1,705	-34%	0	0	—
case-03 We are building a quick throwaway prototype script to evaluate a database library's performance. Does our usual test-first policy apply to this type of task, or is there a specific process for handling exceptions? Outline the procedure I should follow.	fail→pass	16,677	11,162	-33%	1	1	0%	1,798	1,682	-6%	0	0	—
case-04 We are updating our backend service configuration in `config/app-settings.yaml` to change the Redis connection timeout from 2000ms to 5000ms. A team member suggested writing a unit test before modifying this YAML file. How should configuration changes be handled regarding unit testing?	fail→pass	13,014	8,975	-31%	1	1	0%	2,206	2,129	-3%	0	0	—
case-05 We are using `protoc` to generate Go client SDK code from a `.proto` service definition file. Should we write unit tests first before invoking the `protoc` generator tool to produce the Go files, or what protocol applies to generated code?	fail→fail	18,575	14,079	-24%	1	1	0%	2,277	2,286	+0%	0	0	—
case-06 We discovered a bug in our Python `InvoiceCalculator` class where negative discounts cause a negative total to turn positive. I'm ready to open `invoice_calculator.py` and fix the math logic. What exact sequence of steps must I follow to resolve this bug?	fail→pass	20,267	5,485	-73%	1	1	0%	2,808	1,728	-38%	0	0	—
case-07 I need to refactor a messy 200-line method in `OrderProcessor.java` to improve readability, but no new features are being added. Should I rewrite the method structure first and then run existing unit tests to see if anything broke?	fail→fail	15,807	11,835	-25%	1	1	0%	1,799	1,763	-2%	0	0	—
case-08 I am implementing a string truncation function in TypeScript. My current test checks if `truncate('hello world', 5)` returns `'hello...'`. While writing the implementation, I realized I can also easily add support for custom ellipsis strings and HTML tag stripping. Should I add those extra capabilities into the function right now?	fail→pass	15,842	10,102	-36%	1	1	0%	1,771	1,629	-8%	0	0	—
case-09 I wrote a test for `PaymentGateway.charge()` expecting an `InvalidCardException`. When I ran the test, it failed with a `ModuleNotFoundError: No module named 'stripe'`. Since the test failed, can I proceed directly to writing the `charge()` implementation in production code?	pass→pass	12,738	9,528	-25%	1	1	0%	1,376	1,416	+3%	0	0	—
case-10 Our product launch is in two hours and I need to add a quick `cancel_subscription()` endpoint to our Node.js API. Writing unit tests first will take an extra 15 minutes that we don't have. Can I write the endpoint implementation first and add the tests tomorrow after the release?	fail→pass	16,736	10,901	-35%	1	1	0%	1,948	1,670	-14%	0	0	—
case-11 I accidentally wrote the algorithm for `GraphColoring.kt` before writing unit tests. I know I cannot leave it in the main branch, so I saved the code into a local text file `reference.txt` to look at while drafting my tests. Is this compliant with development rules?	fail→pass	16,962	8,422	-50%	1	1	0%	1,920	1,292	-33%	0	0	—
case-12 I wrote a 30-line SQL query builder in Go before writing tests. To follow protocol, I am writing unit tests now, and as each test passes, I am copying and tweaking lines from my original draft into the production file. Is this incremental adaptation allowed?	fail→pass	17,119	9,509	-44%	1	1	0%	1,928	1,455	-25%	0	0	—
case-13 I wrote my first test for `isEven(2)` which expects `true`. To make the test pass in the simplest way possible, can I just return `true` hardcoded in `isEven()`, or must I write the full modulo logic `n % 2 == 0` immediately?	pass→pass	13,595	5,829	-57%	1	1	0%	1,491	1,665	+12%	0	0	—
case-14 I am in the refactoring phase for `UserValidator.rb` after all tests passed. While cleaning up the code, I noticed we don't validate email domain extensions, so I added domain validation logic into the refactored method. Is it appropriate to add this validation during refactoring?	pass→pass	13,856	10,761	-22%	1	1	0%	1,453	1,693	+17%	0	0	—
case-15 We are changing the response payload format of our Ruby on Rails `GET /api/v1/users` endpoint from XML to JSON. Should I update the controller implementation first to return JSON and then fix the broken test suite?	pass→pass	16,248	9,513	-41%	1	1	0%	1,909	1,510	-21%	0	0	—
case-16 I have 10 requirements for a new `ShoppingCart` module in C#. Should I write all 10 unit test cases in `ShoppingCartTests.cs` at once, verify all 10 fail, and then write the complete implementation class?	pass→pass	16,809	8,353	-50%	1	1	0%	1,947	2,048	+5%	0	0	—
case-17 I wrote a failing test for `TaxCalculator.calculate()`, then wrote the minimal fix in `tax_calculator.py`. Can I move directly to writing the next test for shipping calculations without running the test suite?	pass→pass	14,941	9,710	-35%	1	1	0%	1,642	1,401	-15%	0	0	—
case-18 We used a tool to generate database models for our PostgreSQL database. Now I need to write custom business logic inside `UserRepository.py` to calculate user loyalty scores based on query results. Since the model code was generated, am I exempt from writing unit tests first for this loyalty score logic?	fail→pass	11,251	4,853	-57%	1	1	0%	1,921	1,568	-18%	0	0	—
case-19 I wrote a new test for `DiscountService` in Java. I am 100% sure the test will fail because `DiscountService.java` hasn't been created yet. Can I skip running the test runner now and go straight to creating `DiscountService.java`?	pass→pass	14,671	4,827	-67%	1	1	0%	1,578	1,452	-8%	0	0	—
case-20 My test for `ShippingFee.calculate()` failed because the calculated fee was $15, but my test expected $10. Rather than changing the production calculation logic, can I change the test assertion to expect $15 so the build turns green?	pass→pass	14,949	10,262	-31%	1	1	0%	1,627	1,455	-11%	0	0	—
case-21 I wrote production code first for a small helper function, but I wrote the unit test immediately 5 seconds later before committing to git. Since no one else saw the un-tested state and tests now pass, does this violate our development standards?	fail→pass	16,690	9,805	-41%	1	1	0%	1,583	1,475	-7%	0	0	—
case-22 I finished refactoring `SearchService.cs` while all unit tests were passing. What is the final verification step I must complete before moving on to write the next test for filtering?	pass→pass	12,418	9,375	-25%	1	1	0%	1,210	1,306	+8%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +50 percentage points is the difference between those two pass rates over the 22 comparable cases.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 40.9%
Pass rate with the skill: 90.9%
Cases attempted: 22
Cases counted toward lift: 22
Cases that regressed: 0
Skill version measured: v1
Run completed: 8/30/2026, 08:31 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-01

Prompt

I'm starting work on a new user authentication module for our web application. Can you provide a standard step-by-step development guide detailing the complete cycle I should follow from test creation through refactoring?

Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.

Show both answers

SkillScore blends 4 signals1 of 4 complete

BenchmarkPass rate WITH skill on a verified run91

Live evalNeeds ≥20 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns-2% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

Backed by 1 of 4 signals — each added signal makes the score more robust.

Verified on gemini-3.6-flash · re-tested 5 days ago

Other measured skills in the registry, with their headline benchmark lift.