Repository inventory

oimiragieo/agent-studio

Skills indexed from this repository, with install-style signals scoped to the repo.
2 skills12 GitHub stars0 weekly installsJavaScriptGitHubOwner profile

Overview

This skill applies Test-Driven Development to authoring production-ready documentation for agent skills. It teaches a RED-GREEN-REFACTOR cycle for docs: write failing pressure tests, observe baseline behavior, author minimal guidance to fix violations, then iterate to close loopholes. The goal is concise, discoverable guidance that agents reliably follow.

How this skill works

You design pressure scenarios that act as tests for documentation, run them to capture how agents fail, then write focused guidance that addresses those specific failures. After updating the doc, you re-run the scenarios to confirm compliance and refactor to tighten language or close new rationalizations. The process enforces minimal, token-efficient content and clear discovery triggers.

When to use it

  • When creating reusable, agent-facing guidance that must be reliably followed
  • When a technique or pattern will be referenced across projects or many agents
  • When agents behave inconsistently without explicit documentation (race conditions, flaky reasoning)
  • When you need documentation that must be discoverable and token-efficient for future agents

Best practices

  • Always run a baseline pressure scenario before writing documentation (watch it fail)
  • Write minimal, targeted guidance that specifically addresses observed failures
  • Start the skills description with clear triggering conditions—don’t summarize the workflow there
  • Use concise naming and keyword coverage so future agents can find the guidance quickly
  • Move heavy details into helper files or tooling and cross-reference rather than repeating content

Example use cases

  • Documenting a debugging pattern so subagents reproduce a consistent diagnostic sequence
  • Capturing a repeatable test scenario for timing-dependent behaviors (race conditions, flaky tests)
  • Authoring a concise reference for a reusable tool or script that many agents will invoke
  • Creating discovery-focused descriptions so agents load the correct guidance only when triggers match

FAQ

You must understand the core TDD cycle (RED-GREEN-REFACTOR) before using this approach.

How long should a skill be?

Keep frequently-loaded guidance under ~200 words; other reusable guidance should remain concise, ideally under ~500 words to save tokens and speed discovery.

2 skills

More from this maintainer
Other repositories and skills published under the same GitHub owner.
Skills library
Jump back to the full directory or explore grouped topics.
Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational