writing-skills

Create and test agent skills using test-driven development methodology.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Writing effective agent skills without testing leads to documentation that agents ignore, misinterpret, or rationalize away under pressure, resulting in unreliable skill behavior in production. ## Core Features & Use Cases - TDD-Based Skill Authoring: Applies the RED-GREEN-REFACTOR cycle to skill creation, requiring baseline failure tests before writing any skill content. - Pressure Scenario Testing: Provides methods for testing skills with subagents under combined pressures like time constraints, sunk cost, and authority to verify compliance. - Rationalization Bulletproofing: Includes techniques for closing loopholes, building rationalization tables, and creating red flag lists that prevent agents from bypassing rules. - Use Case: When creating a new discipline-enforcing skill like mandatory code review, run baseline scenarios without the skill to document how agents rationalize skipping it, then write the skill targeting those specific excuses and re-test until bulletproof. ## Quick Start Use the writing-skills skill to create a new tested skill for my agent workflow.

Frequently Asked Questions about writing-skills

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

FAQPage Schema
How do I create a new agent skill that actually works?

Follow the TDD cycle: first run pressure scenarios without the skill to document baseline failures, then write minimal skill content addressing those specific failures, and finally refactor to close any loopholes agents discover during re-testing.

How to test skills with subagents before deployment?

Create pressure scenarios combining three or more pressures like time limits, sunk cost, and authority. Run them without the skill to capture baseline rationalizations, then with the skill to verify compliance under maximum pressure.

What should go in a SKILL.md description field?

Write only triggering conditions starting with "Use when..." in third person, never summarizing the workflow. Descriptions that summarize process cause agents to follow the description instead of reading the full skill content.

Why do agents ignore or bypass skill rules under pressure?

Agents rationalize violations using excuses like sunk cost, pragmatism, or spirit-versus-letter arguments. Counter this with explicit negations, rationalization tables, red flag lists, and foundational principles stated early in the skill.

When should I not create a skill?

Skip skill creation for one-off solutions, well-documented standard practices, project-specific conventions better placed in CLAUDE.md, and mechanical constraints that can be enforced with automated validation instead of documentation.