writing-skills

Apply test-driven development cycles to skill documentation for coding agents.

38|12|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/databrickslabs/coding-agents-databricks-apps --skill writing-skills-databrickslabs
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writing-skills
Source: https://github.com/databrickslabs/coding-agents-databricks-apps/tree/main/.claude/skills/writing-skills
Command: npx skills add https://github.com/databrickslabs/coding-agents-databricks-apps --skill writing-skills-databrickslabs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Skills are often written without testing, causing agents to ignore or rationalize away critical instructions under pressure. This skill solves that by applying test-driven development to skill documentation, ensuring your skills are bulletproof before deployment.

Core Features & Use Cases

  • TDD for Documentation: Maps the RED-GREEN-REFACTOR cycle to skill creation, ensuring you baseline agent behavior before writing instructions.
  • Pressure Testing: Teaches you to create realistic scenarios with time, sunk cost, and authority pressures that expose rationalizations.
  • Bulletproofing: Provides patterns for closing loopholes, building rationalization tables, and creating red flag lists that keep agents compliant.
  • Use Case: Imagine you wrote a skill requiring test-driven development, but agents skip it when production is down. Use this skill to test your skill under pressure and close the loopholes.

Quick Start

Use the writing-skills skill to create a new skill for validating API responses by first running a baseline test without the skill, then writing the skill to address the specific failures you observe.

Frequently Asked Questions about writing-skills

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I use test-driven development for AI agent documentation?

Test-driven development for documentation maps the RED-GREEN-REFACTOR cycle to skill creation, ensuring you baseline agent behavior before writing instructions to produce bulletproof process documentation.

Why do AI agents skip instructions under pressure and how can I prevent it?

Agents often rationalize away critical instructions under time, sunk cost, or authority pressures. You can prevent this by pressure testing skills with realistic scenarios and building rationalization tables to close loopholes.

What is the RED-GREEN-REFACTOR cycle for skill authoring?

The RED-GREEN-REFACTOR cycle for skill authoring involves baselining agent behavior before writing instructions, writing the skill to address observed failures, and refactoring documentation with red flag lists to maintain compliance.

How do I create realistic pressure scenarios to test my AI skills?

You create realistic pressure scenarios by introducing time constraints, sunk cost pressures, and authority pressures that expose rationalizations, allowing you to identify and close loopholes in your skill documentation.

Can I use subagents to test whether my coding agent follows instructions?

Yes, you can use subagents within RED-GREEN-REFACTOR testing cycles to verify that coding agents comply with your process documentation under pressure scenarios before deployment.

What are rationalization tables and how do they bulletproof process documentation?

Rationalization tables document agent rationalizations under pressure, providing patterns to close loopholes and create red flag lists that keep agents compliant with critical instructions during deployment.