Skill

tdd

Use when you want to write tests first, enforce test-driven development, or add test coverage — 'write tests for this', 'TDD this feature', 'add test coverage'. Strict RED-GREEN-REFACTOR discipline.

From dp-cto

Install

Run in your terminal

npx claudepluginhub raisedadead/dotplugins --plugin dp-cto

Tool Access

This skill uses the workspace's default tool permissions.

Supporting Assets

View in Repository

references/anti-rationalization.md

references/good-tests.md

references/implementation-examples.md

references/red-flags.md

references/test-examples.md

references/verification-checklist.md

Skill Content

Similar Skills

agent-harness-construction

Designs and optimizes AI agent action spaces, tool definitions, observation formats, error recovery, and context for higher task completion rates.

ecc

145.8k

agent-introspection-debugging

Implements structured self-debugging workflow for AI agent failures: capture errors, diagnose patterns like loops or context overflow, apply contained recoveries, and generate introspection reports.

ecc

145.8k

agent-eval

Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.

ecc

145.8k

Stats

Parent Repo Stars3

Parent Repo Forks0

Last CommitApr 1, 2026

Actions

View Source View Plugin View on GitHub View README

dp-cto:tdd — TDD Enforcement

Read ${CLAUDE_PLUGIN_ROOT}/shared/cto-boundaries.md

The Iron Law

NO PRODUCTION CODE WITHOUT A FAILING TEST FIRST

Wrote code before the test? Delete it. Not "set aside." Not "use as reference." Delete. Implement fresh from the failing test.

Protocol

1. Gather Test Target

Collect from the user:

Feature or behavior to test
Target file paths (resolve concrete paths — never placeholders)
Verification command (test runner, specific test file)
Whether this is a new feature or a bug fix (determines RED-GREEN-REFACTOR vs bug fix protocol)

If the user's description is vague, ask clarifying questions before dispatch. One question per message.

2. Dispatch Tester Agent

Dispatch tester via the Agent tool with subagent_type: "dp-cto:tester".

The agent prompt must include:

The test target (feature description or bug to reproduce)
Concrete file paths (source and test files)
The verification command to run tests
File scope constraints (what the agent may and may not modify)

The tester agent follows strict RED-GREEN-REFACTOR discipline: write one failing test, verify RED, write minimal code to pass, verify GREEN, refactor. The agent definition enforces this cycle.

3. Validate Receipt

On agent completion, extract the structured receipt from  markers. Validate:

tests_written — list of tests created
tests_watched_fail — every test must have been watched failing (RED verified)
tests_passing — all tests green after implementation
verification — command, exit code, and output preview

If tests_watched_fail is empty or does not match tests_written, reject the receipt. A test never seen failing is a test you do not trust.

4. Report Results

Present the tester's results to the user: tests written, RED/GREEN verification evidence, and any unresolved issues from the receipt.

Read ${CLAUDE_SKILL_DIR}/references/test-examples.md before proceeding. Read ${CLAUDE_SKILL_DIR}/references/implementation-examples.md before proceeding. Read ${CLAUDE_SKILL_DIR}/references/anti-rationalization.md before proceeding. Read ${CLAUDE_SKILL_DIR}/references/red-flags.md before proceeding. Read ${CLAUDE_SKILL_DIR}/references/good-tests.md before proceeding. Read ${CLAUDE_SKILL_DIR}/references/verification-checklist.md before proceeding.

NEVER

NEVER write tests or production code inline — always dispatch the tester agent
NEVER accept a receipt where tests_watched_fail is empty
NEVER skip receipt validation — verify RED-GREEN-REFACTOR evidence
NEVER auto-invoke follow-up skills — present results and let the user decide
NEVER rationalize skipping TDD — check the anti-rationalization table