Viget Agent Skills

Warning

This is a public repository. Skills should be general-use or contain only publicly available Viget information. Do not commit proprietary or internal-only content.

A collection of skills for AI coding agents. Works with Claude Code, Cursor, Codex, GitHub Copilot, Gemini CLI, more! Based on the agentskills.io standard.

Skills We Like

Development

• superpowers — A complete software development workflow for coding agents, with composable skills for design refinement, planning, and test-driven implementation.
• compound-engineering-plugin — A Skills plugin emphasizing planning, review, and knowledge codification to make subsequent engineering work easier.

Install

Claude Web & Desktop

Upload skills directly in the Claude app — no CLI needed. See Using Skills in Claude Web & Desktop.

Any agent (recommended)

npx skills add vigetlabs/viget-agent-skills

skills by Vercel is a CLI for installing and managing reusable skills across 37+ AI coding agents. Requires Node.js and npm — npx runs the package without a global install.

The CLI auto-detects your installed agents and symlinks the skills into each one's expected directory. Options:

# Install a single skill
npx skills add vigetlabs/viget-agent-skills --skill skill-name

# Install globally (available in all projects)
npx skills add vigetlabs/viget-agent-skills -g

# Target a specific agent
npx skills add vigetlabs/viget-agent-skills -a claude-code

Claude Code

/install-plugin vigetlabs/viget-agent-skills

Manual

Clone the repo and symlink the skills you want:

git clone https://github.com/vigetlabs/viget-agent-skills.git ~/.agent-skills/viget-agent-skills

# Claude Code
ln -s ~/.agent-skills/viget-agent-skills/skills/* ~/.claude/skills/

# Cursor
ln -s ~/.agent-skills/viget-agent-skills/skills/* ~/.cursor/skills/

# Codex
ln -s ~/.agent-skills/viget-agent-skills/skills/* ~/.codex/skills/

Symlinks keep your skills in sync with git pull.

---

Contributing a Skill

Quick start

1. Create a new directory under skills/ with your skill name
2. Add a SKILL.md file with frontmatter and instructions
3. Open a PR

skills/
└── your-skill-name/
    ├── SKILL.md            # Required
    ├── scripts/            # Optional: executable code
    ├── references/         # Optional: supplementary docs
    └── assets/             # Optional: templates, schemas, etc.

SKILL.md format

Every skill needs a SKILL.md with YAML frontmatter and a Markdown body:

---
name: your-skill-name
description: >-
  What this skill does. Use when [specific trigger conditions]. Handles
  [keywords agents will match on].
---

The body contains the actual instructions the agent will follow. Write it as clear, step-by-step guidance with examples and edge cases.

Frontmatter fields

Field	Required	Notes
name	Yes	Lowercase, hyphens only, max 64 chars. Must match the directory name.
description	Yes	Max 1024 chars. Describe what it does and when to use it.
license	No	License name or reference to a bundled file.
compatibility	No	Environment requirements (e.g. "Requires docker").
allowed-tools	No	Space-delimited list of pre-approved tools. Experimental.
metadata	No	Arbitrary key-value pairs (author, version, etc.).

Writing good descriptions

The description is the only thing agents read at startup to decide if your skill is relevant. This makes it the most important field.

Do:

• Describe what the skill does and when to trigger it
• Include specific keywords agents will match on
• Start with a verb ("Extracts text from PDFs", "Runs database migrations")

Don't:

• Summarize the workflow. Agents will follow the summary instead of reading the full SKILL.md.
• Be vague ("Helps with files")

Managing context budget

Skills use progressive disclosure to stay lightweight:

1. Description (~100 tokens) -- loaded at startup for all installed skills
2. SKILL.md body (<5000 tokens) -- loaded when the skill activates
3. References/scripts -- loaded only when the skill explicitly asks for them

Rules of thumb:

• Keep SKILL.md under 500 lines
• Move detailed reference material to references/
• Keep file references one level deep (don't chain references to references)
• Scripts run externally and cost zero context tokens

Optional directories

scripts/ -- Executable code the agent can run.

• Use stderr for human-readable output, stdout for structured JSON
• Include set -e for fail-fast
• Add cleanup traps for temporary resources
• Document dependencies at the top of the script

references/ -- Supplementary docs loaded on demand.

• Keep each file focused on one topic
• Name files descriptively (api-endpoints.md, not ref1.md)
• Reference them from SKILL.md: See [API docs](references/api-endpoints.md)

assets/ -- Static resources: templates, schemas, config files, images.

Example skill

skills/
└── lint-fix/
    ├── SKILL.md
    └── references/
        └── eslint-rules.md

# skills/lint-fix/SKILL.md
---
name: lint-fix
description: >-
  Runs linting and auto-fixes code style issues. Use when the user asks to
  lint, fix formatting, or clean up code style. Supports ESLint, Prettier,
  and Stylelint.
---

# Lint and Fix

## Process

1. Detect which linters are configured in the project (check for
   `.eslintrc*`, `.prettierrc*`, `.stylelintrc*`, `package.json` config)
2. Run the appropriate fix command
3. Report what changed

## Common edge cases

- If no linter config exists, ask the user before installing one
- If multiple linters conflict, prefer the project's existing config
- See [ESLint rule reference](references/eslint-rules.md) for rule-specific
  guidance

Checklist before opening a PR

• Directory name matches the name field in frontmatter
• Description says what it does and when to trigger (not how it works)
• SKILL.md is under 500 lines
• No agent-specific tool names in the SKILL.md body (use conditional language instead)
• Scripts are self-contained with documented dependencies
• Tested with at least one agent