Create agent skills via RED-GREEN-REFACTOR with SKILL.md and pressure tests.
npx skills add https://github.com/Wcof/PMSkill --skill pm-skillauthor
Please help me install this Agent Skill. Skill: pm-skillauthor Source: https://github.com/Wcof/PMSkill/tree/main/skills/discovery/pm-skillauthor Command: npx skills add https://github.com/Wcof/PMSkill --skill pm-skillauthor
This Skill helps teams create new AI skills without guessing what good documentation looks like. It replaces vague skill authoring with a test-driven workflow that first observes how an agent fails without the skill, then writes targeted instructions to correct those failures, and finally hardens the result with pressure testing.
Ask the agent to use pm-skillauthor to create a new skill for a specific capability by first recording baseline failures, then drafting SKILL.md against those failures, and finally running refactor pressure tests.
High-intent search queries and answers about installing and using this skill.
Agent skills can be authored using a RED-GREEN-REFACTOR workflow that records baseline agent failures, converts them into targeted SKILL.md instructions, and hardens the result with refactor pressure testing.
TDD for skill authoring replaces vague documentation by first observing how an agent fails without the skill, then writing targeted instructions to correct those specific failures, and finally hardening the result with pressure testing.
Compliant Anthropic skills require specific frontmatter, structured sections, failure modes, anti-pattern blacklists, references, and test-prompts artifacts to ensure proper routing triggers and failure handling for reusable PM workflows.
Pressure-test hardening checks whether a drafted skill holds up under time pressure, authority pressure, and sunk-cost bias, ensuring the final SKILL.md documentation remains robust in real-world product team scenarios.
This process applies to reusable skill development for product management and agent workflows, helping teams add capabilities like interview note structuring or experiment analysis through a disciplined authoring process.
Recording baseline failures first captures concrete agent mistakes, which are then converted into targeted SKILL.md instructions, ensuring the documentation corrects actual problems rather than producing generic instructions.
nimeshgurungApply the RED-GREEN-REFACTOR cycle to create and refine AI agent Skills.
nateaValidate n8n webhook, schedule, polling, and event triggers.
obraAutomate RED-GREEN-REFACTOR testing of skills with subagents.
chuangkevinAuthor and verify agent Skills using red-green-refactor pressure scenarios.
