What Are the Modular Skills Provided by Ponytail? A Complete Guide

Ponytail provides six modular skills—ponytail, ponytail-review, ponytail-audit, ponytail-debt, ponytail-gain, and ponytail-help—each defined in self-contained SKILL.md files under the skills/ directory and triggered via slash commands.

Ponytail is an open-source lazy-developer assistant that enforces minimalism through a modular skill architecture. Each modular skill is a self-contained prompt definition stored in its own subdirectory, allowing the runtime to discover and load capabilities without code changes. This design separates behavior definitions from execution logic, making the system extensible by simply adding new folders to the repository.

Overview of Ponytail's Modular Skill System

The modular skills provided by Ponytail follow a strict file-based convention. Every skill resides in skills/<name>/SKILL.md and includes YAML front-matter (name, description, argument-hint, license) followed by the prompt template. The core plugin located at hooks/ponytail-instructions.js walks the skills/ directory at runtime, automatically registering any valid SKILL.md as a slash command.

This architecture ensures that skills remain documentation-first: the SKILL.md serves as both the user-facing documentation and the runtime prompt source. Build scripts such as scripts/build-openclaw-skills.js and scripts/publish-openclaw-skills.js copy these files verbatim into the OpenClaw plugin bundle, guaranteeing zero drift between documentation and execution.

The Six Core Modular Skills

Ponytail ships with six built-in skills, each targeting a specific phase of the development lifecycle.

The Core ponytail Skill

The foundational skill lives in skills/ponytail/SKILL.md and implements the lazy-developer mode. Triggered via /ponytail (or variants like /ponytail lite, /ponytail full, and /ponytail ultra), this skill enforces a strict "ladder" of priorities: YAGNI → stdlib → native → one-line → minimal.

The intensity levels control how aggressively the assistant challenges over-engineering:

• lite: Gentle suggestions for simplification
• full: Strict enforcement of minimal solutions
• ultra: Maximum bias against any abstraction

Code Review with ponytail-review

Located at skills/ponytail-review/SKILL.md, this skill performs targeted diff analysis. When invoked with /ponytail-review, it scans the provided diff and lists every instance of over-engineering—including dead code, unnecessary dependencies, and YAGNI violations—in a compact one-line format per finding.

Repository Audits with ponytail-audit

The skills/ponytail-audit/SKILL.md skill scales the review concept to the entire codebase. Using /ponytail-audit, it conducts a repo-wide search for the same anti-patterns as ponytail-review, ranking findings by impact to prioritize refactoring efforts.

Technical Debt Tracking with ponytail-debt

Defined in skills/ponytail-debt/SKILL.md, this skill harvests every ponytail: comment left in the source code. These comments represent shortcuts, deferrals, or acknowledged compromises. The skill aggregates them into a ledger, ensuring technical debt remains visible and trackable rather than silently rotting in the codebase.

Impact Metrics with ponytail-gain

The skills/ponytail-gain/SKILL.md skill provides quantitative justification for minimalism. Triggered by /ponytail-gain, it displays a scoreboard of measured impact based on benchmark medians, including lines of code saved, cost reduction, and speed improvements achieved through Ponytail's lazy-development approach.

Help and Documentation with ponytail-help

Finally, skills/ponytail-help/SKILL.md serves as the system's documentation layer. The /ponytail-help command displays a quick-reference card listing all available modes, intensity levels, and skill triggers, making the modular skills discoverable without leaving the chat interface.

How the Skill System Works

The modularity relies on three architectural components that separate definition from execution.

Discovery via Filesystem

The runtime loader in hooks/ponytail-instructions.js recursively walks the skills/ directory, identifying any folder containing a SKILL.md file. It parses the YAML front-matter to register the command name and description, then loads the markdown body as the prompt template.

Single Source of Truth

Each SKILL.md functions as the authoritative definition. The scripts/build-openclaw-skills.js script copies these files directly into the plugin bundle without transformation, while scripts/publish-openclaw-skills.js handles distribution. This ensures that editing skills/ponytail-audit/SKILL.md immediately updates both documentation and runtime behavior.

Uniform Interface

All skills share an identical front-matter schema:

---
name: skill-name
description: Human-readable purpose
argument-hint: Usage instructions
license: MIT
---

This uniformity allows hooks/ponytail-instructions.js to treat heterogeneous capabilities—ranging from diff analysis to debt tracking—as interchangeable modules.

Adding Custom Modular Skills

Extending Ponytail requires no code changes. To create a custom skill, create a new directory under skills/ with a valid SKILL.md:

mkdir -p skills/ponytail-custom
cat > skills/ponytail-custom/SKILL.md <<'EOF'
---
name: ponytail-custom
description: Validates variable naming conventions against Ponytail minimalism standards
argument-hint: "<filepath>"
license: MIT
---
Review the provided file for verbose variable names. Suggest shorter alternatives that maintain clarity while reducing character count. Prefer single-letter variables only in loop contexts.
EOF

After reloading the plugin, the skill automatically registers as /ponytail-custom and appears in /ponytail-help output.

Summary

• Ponytail provides six built-in modular skills covering core assistance, code review, auditing, debt tracking, metrics, and documentation.
• Each skill is a self-contained SKILL.md file under skills/<name>/, using standardized YAML front-matter.
• Automatic discovery occurs through hooks/ponytail-instructions.js, which walks the filesystem at runtime.
• Zero-drift deployment is ensured by build scripts (scripts/build-openclaw-skills.js) that copy skill definitions verbatim into the runtime bundle.
• Custom skills require only a new folder and markdown file, immediately available as slash commands without recompilation.

Frequently Asked Questions

How do I invoke a specific modular skill in Ponytail?

Use the slash command corresponding to the skill name defined in the SKILL.md front-matter. For example, type /ponytail-review to trigger the review skill, /ponytail-audit for repository-wide analysis, or /ponytail-debt to view the technical debt ledger. The /ponytail-help command lists all available triggers.

What is the difference between ponytail-review and ponytail-audit?

ponytail-review (skills/ponytail-review/SKILL.md) analyzes a specific diff or code block for over-engineering, providing immediate feedback on changes. ponytail-audit (skills/ponytail-audit/SKILL.md) performs a comprehensive scan of the entire repository, ranking violations by impact to guide large-scale refactoring priorities.

Can I create my own modular skills for Ponytail?

Yes. Create a new directory under skills/ containing a SKILL.md file with valid YAML front-matter (name, description, argument-hint, license). The hooks/ponytail-instructions.js loader automatically discovers and registers new skills on the next plugin reload, making them available as slash commands without modifying any JavaScript code.

Where are the skill definitions stored in the Ponytail repository?

Skill definitions reside in skills/<skill-name>/SKILL.md paths. Core skills include skills/ponytail/SKILL.md (main assistant), skills/ponytail-review/SKILL.md (diff analysis), skills/ponytail-audit/SKILL.md (repo auditing), skills/ponytail-debt/SKILL.md (debt tracking), skills/ponytail-gain/SKILL.md (impact metrics), and skills/ponytail-help/SKILL.md (documentation).

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}