Agent Meta Scaffolder

Agent Meta-Scaffolder Skill

This skill empowers AI agents to autonomously expand their capabilities by creating and managing new skills within the .agent/skills directory.

Core Capabilities

1. Skill Generation: Automatically generate SKILL.md files with correct YAML frontmatter and structure based on a given domain.
2. Context Extraction: Scan the codebase, documentation, and external resources (using search_web) to distill "procedural knowledge" into reusable skills.
3. Standards Compliance: Ensure all generated skills follow the shared specifications of OpenAI, Anthropic, and other major agentic platforms.
4. Discovery & Linking: Identify gaps in the current agent capabilities and suggest new skills to fill them.

Skill Creation Workflow

1. Identify Need: Determine if a new task or domain requires a dedicated skill (e.g., "Add support for a new game engine protocol").
2. Gather Knowledge: Use run_command, view_file, and search_web to collect best practices.
3. Draft SKILL.md: Use the predefined template to create a new folder and SKILL.md.
4. Verify & Refine: Run lint checks on the new markdown and verify the instructions are actionable for other agent instances.

Skill Template (Reference)

---
name: skill-name
description: human-readable description
version: 1.0.0
---

# Skill Title

## Capabilities

1. ...

## Guidance

- ...

## Code Snippets

...

Maintenance Tasks

• Audit existing skills: Check for outdated instructions or broken links.
• De-duplicate: Merge skills that have overlapping responsibilities.
• Update metadata: Ensure versioning and descriptions accurately reflect the skill's utility.

Future Plans (Self-Evolution)

• Automated testing of skills in sandboxed environments.
• Usage tracking to prioritize which skills to keep and which to archieve.

Converted and distributed by TomeVault — claim your Tome and manage your conversions.