agentsclimarketplace

AGENTSKILL

Skill MYMDO/AGENTSKILL

Complete guide to minimal universal configuration files compatible with all major AI coding agents.

Install
npx -y skills add MYMDO/AGENTSKILL

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.0 KB, 905 tokens by cl100k_base, as published. Nobody here has run it

AGENTSKILL.md - Skill Creator and Config Generator

Definitive workflow for AI agents updating this repository or using it to create downstream agent configuration.

Purpose

This repository maintains a professional starter kit for AGENTS.md and SKILL.md files compatible with major AI coding agents and the Agent Skills v1.0.0 model.

Root files describe this repository. Copy-ready files live in templates/. Filled examples live in examples/.

Five-Phase Workflow

Phase 1: Research

Before writing or changing configuration, inspect the target project:

  1. Read README files and root manifests.
  2. Check existing agent files: AGENTS.md, CLAUDE.md, .cursor/, .github/, .pi/, .opencode/.
  3. Identify install, dev, test, build, lint, format, and typecheck commands from executable sources.
  4. Inspect CI workflows and pre-commit hooks.
  5. Read representative source files for architecture, entrypoints, and conventions.
  6. Note repo-specific gotchas an agent would likely miss.

Prefer executable truth over prose. If scripts and docs disagree, trust scripts and update stale docs.

Phase 2: Create or Update AGENTS.md

Use templates/AGENTS.template.md as the canonical downstream template.

Quality rules:

  • Stay under 200 lines.
  • Include exact commands.
  • Include architecture boundaries and gotchas.
  • Keep generic advice out.
  • Use {{UPPER_SNAKE_CASE}} placeholders only in templates, not filled examples.
  • Document agent adapters only where relevant.

Phase 3: Create or Update SKILL.md

Use templates/SKILL.template.md as the canonical skill template.

Required frontmatter:

---
name: skill-name
description: What the skill does. Use when specific trigger scenarios happen.
license: MIT
---

Validation checklist:

  • name is 1-64 characters, lowercase, hyphenated, no leading/trailing/consecutive hyphens.
  • description is 1-1024 characters.
  • description includes both what the skill does and when to trigger it.
  • Steps are executable without guessing.
  • Examples are concrete.
  • Optional scripts/, references/, and assets/ are referenced only when useful.

Phase 4: Document for Humans

Use README for the shortest successful path:

  • What the repo is.
  • How to copy or scaffold templates.
  • How compatibility works.
  • Where examples and docs live.
  • How to validate.

Use docs/ for deeper material:

  • docs/compatibility.md
  • docs/quality-rubric.md
  • docs/anti-patterns.md
  • docs/troubleshooting.md

Phase 5: Verify

Run:

python3 scripts/validate.py

Check:

  • Root AGENTS.md is under 200 lines.
  • Template skill frontmatter is valid.
  • Example skills are valid.
  • Examples have no unresolved placeholders.
  • Internal Markdown links resolve.
  • agentskill.json references existing files and directories.
  • Terminology is consistent across README, docs, templates, and root skill.

File Relationships

AGENTS.md                 -> Context for this repository
SKILL.md                  -> Skill for setting up downstream agent config
templates/AGENTS.template.md
                           -> Copy-ready project-context template
templates/SKILL.template.md
                           -> Copy-ready skill template
examples/*                -> Filled stack-specific examples
docs/*                    -> Human reference material
scripts/validate.py       -> Repository validation
scripts/scaffold-agent-config.sh
                           -> Template scaffolding helper
agentskill.json           -> Machine-readable manifest

Compatibility Model

Use precise support language:

  • Native: the agent reads the standard location directly.
  • Adapter: a symlink, copy, or agent-specific directory is required.
  • Manual: the user must point the agent at the file or content.

Do not describe adapter-based support as native.

Resources

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.