Turn real workflows into reusable Agent capabilities.
An independent skill catalog for Codex, Claude Code, and compatible Agents.
Every skill has its own repository, installation guide, use cases, and validation record.
Quick start · Skill catalog · Compatibility · Contact
| 🔎 Easier to discover | 📦 Easier to install | ✅ Easier to validate |
|---|---|---|
| Search for a problem and land on the relevant skill directly. | Every repository has its own README, version, Issues, and install path. | Structure, scripts, and compatibility evidence stay focused. |
Choose a skill and give its installation guide to your Agent:
Install <skill-name>: https://raw.githubusercontent.com/Fable-Forge/<skill-name>/main/docs/install.md
Replace <skill-name> with the skill you want to install, such as sol-luna-local-parallel, product-teardown, or sprite-atlas-slicer. Your Agent will clone it into the appropriate directory, so you do not need to manage the path yourself.
Or run:
npx skills add Fable-Forge/<skill-name>This catalog contains 20 independent skills. Structural validation, installation visibility, and real triggering are recorded separately and do not substitute for one another.
| Skill | Outcome | Maturity |
|---|---|---|
cocos-creator-mobile-game |
A production workflow for Cocos Creator 3.x mobile games. | beta |
game-entity-movement-presentation |
Coordinate movement and presentation for gameplay entities. | beta |
game-event-director |
Design and implement game events, encounters, triggers, and scripted beats. | beta |
game-ui-panels |
Wire game HUDs, menus, dialogs, and panels to real state and interactions. | beta |
godot-narrative-workflow |
A docs-first production workflow for narrative and low-art Godot projects. | beta |
narrative-game-prototype-planner |
Turn narrative game notes into testable mechanics, states, and prototype plans. | beta |
scene-anchor-system |
Create consistent scene spaces, camera anchors, and composition constraints. | beta |
| Skill | Outcome | Maturity |
|---|---|---|
product-image-prompt-iteration |
Systematically iterate prompts for product renders, three-view sheets, and hero images. | beta |
screen2layers |
Reconstruct flat UI screenshots into movable slices and Figma-ready layered kits. | beta |
sprite-atlas-slicer |
Detect, split, and verify individual assets from transparent sprite atlases. | beta |
| Skill | Outcome | Maturity |
|---|---|---|
agent-skill-migration |
Safely inventory, migrate, sync, and verify skills across agent runtimes. | beta |
agents-md-verification-map |
Build repository-true, tiered verification maps for AGENTS.md. | beta |
reporting-discipline |
Turn long evidence streams into concise, honest, verifiable reports. | beta |
sol-luna-local-parallel |
Coordinate Sol-led, Luna-parallel work for non-overlapping local tasks. | experimental |
unfinished-work-audit |
Audit local Codex and Claude sessions against repositories to find genuinely unfinished work. | experimental |
windows-ai-cli-recovery |
Diagnose and recover Codex, Claude Code, and npm AI CLI failures on Windows. | beta |
| Skill | Outcome | Maturity |
|---|---|---|
obsidian-article-writeback |
Turn project and research insights into deduplicated Obsidian knowledge articles. | beta |
offline-ux-kit |
Add consent-aware local UX evidence, categorized feedback, and self-contained reports to offline web prototypes. | beta |
product-teardown |
Evidence-led product, game, and methodology teardown for actionable transfer. | beta |
xiaohongshu-copywriter |
Create, rewrite, diagnose, and package natural Chinese content for Xiaohongshu. | beta |
beta: structure and local installation paths are validated; broader cross-machine usage still needs feedback.experimental: depends on particular runtime capabilities or local session formats; inspect the environment before use.- Each independent repository README is the authority for current support.
- 📧 Email: 53815263@qq.com
- 🐛 Bugs and feature requests: GitHub Issues
- 💬 Usage questions and business collaboration: GitHub Discussions or email
You do not need a complete proposal before reaching out. A real workflow, problem, or requirement is enough to begin evaluating whether an Agent can solve it and how.
MIT © 2026 FableForge