GSD Skill Creator — For AI Agents & LLMs
This page provides structured entry points for AI agents, LLMs, and automated tools to discover and consume documentation for GSD Skill Creator.
Machine-readable documentation:
llms.txt— Curated index with descriptions and raw GitHub markdown linksllms-full.txt— All documentation concatenated into a single file (~200KB)
These files follow the llms.txt standard and are served at the site root for automatic discovery.
Quick Install
# Install via npx (recommended)
npx get-shit-done-cc@latest
# Or clone from source
git clone https://github.com/Tibsfox/gsd-skill-creator.git
cd gsd-skill-creator
npm install && npm run build && npm linkRepository: github.com/Tibsfox/gsd-skill-creator
What Is This Project?
GSD Skill Creator is an adaptive learning layer for Claude Code, built as an extension to GSD (Get Shit Done).
Key Concepts
| Concept | Description |
|---|---|
| Skills | Compact 300-500 token markdown files that encode patterns, best practices, and domain knowledge. Auto-loaded into Claude Code sessions based on context. |
| GSD Workflow | Structured development lifecycle: Discuss → Plan → Execute → Verify. Externalizes project state into markdown files (PROJECT.md, ROADMAP.md, STATE.md). |
| Bounded Learning | Safety guardrails for skill evolution: max 20% change per refinement, 7-day cooldowns, user confirmation required. Prevents runaway self-modification. |
| Pattern Detection | Observes recurring tool sequences and user corrections across sessions, then suggests new skills from detected patterns. |
| Token Budget | 6-stage pipeline allocating 2-5% of context window to skills, with priority scoring and cache-aware ordering. |
| Agent Teams | Multi-agent coordination where skills compose into specialized agents from co-activation patterns. |
Documentation Structure
The documentation is organized into the same sections as llms.txt:
Getting Started
Getting Started Guide — Installation, prerequisites, first skill creation
How It Works — Observation-to-skill pipeline
Core Concepts — Skills, patterns, bounded learning, token budgets
Configuration — All configuration options
Links
- GitHub Repository
- Documentation Hub (human-friendly browsing)
llms.txt(machine-readable index)llms-full.txt(full docs, single file)- README.md (project overview)
- CLAUDE.md (Claude Code integration guide)