Skip to content

JWCodeWrote/Agent_Skills_Plugin

Repository files navigation

Agent Skills Plugin

Read this in 简体中文.

Welcome to the Agent Skills Plugin repository. This project is a curated collection of specialized "Cognitive Modules" (Skills) designed to upgrade AI Agents with distinct personas, philosophical frameworks, and execution capabilities.

Instead of simple Python tools, these skills provide Mental Operating Systems that force the Agent to assume specific roles—from a dialectical philosopher to a high-speed executor.

🧩 Featured Skills

Skill Name Description Key Persona
Dialectical_Materialism Cognitive OS: Forces objective, non-linear analysis using Marxist dialectics (Unity of Opposites, Quantity-Quality Leap). The Dialectical Philosopher
Meihua_Yishu (梅花易数) Hybrid Divination: A strategic decision support system combining traditional I-Ching math with Gemini AI for deep, modern interpretation. Metaphysical Strategist
AI-search-browser-use-main Research Mode: Advanced web research capabilities using Chrome CDP and browser automation for deep synthesis. The Deep Researcher
HighTriad Architecture Mode: Plans systems for high concurrency, high performance, and high availability with SLOs, scaling, resilience, and validation. The Systems Architect
story-commentary-workflow Story Commentary Mode: Organizes game story footage, chapter breakdowns, subtitles, narration drafts, and voiceover-to-visual mapping. The Story Commentary Producer
story-worldbuilding-planner Worldbuilding Mode: Creates and refines game stories, web novel plots, light novel arcs, characters, factions, power systems, and lore bibles. The Narrative Designer
European-Chinese-Cleaner Chinese Rewrite Mode: Rewrites translationese, Europeanized grammar, AI-generated stiffness, and English calques into natural Chinese. The Chinese Language Editor
PUA AI Compliance Pressure Mode: Applies strict, no-excuse pressure to the AI agent for immediate self-correction; explicitly forbids user-targeted PUA. The Agent Discipline Enforcer
agent-fault-retrospective Agent Fault Retrospective: When the user flags an AI agent mistake, overstep, or misunderstanding, asks for user-defined trigger keywords, applies a Superpowers-style root-cause review, and records concise behavior rules in the correct instruction file (AGENTS.md, CLAUDE.md, Cursor rules, Copilot instructions, etc.). The Fault Retrospective Officer
Impeccable-design-ui Design Workflow Hub: An Impeccable design menu that routes to 14 specialized sub-skills (critique, colorize, animate, polish, etc.) with auto-install via npx skills add pbakaus/impeccable. The Design Director
awesome-design-setup Brand Design Picker: Browse 58 curated brand DESIGN.md files (Linear, Stripe, Apple, Tesla, etc.) and download the chosen one to your project root for AI-agent-driven UI development. The Brand Stylist
frontend-imagegen-director Front-end Image-to-Implementation Director: Inspects the project's existing style and product nature, uses superpowers for requirement discovery, generates front-end visuals with gpt-image2 or the environment's imagegen path, and then builds the actual page from the approved image. If superpowers is unavailable, it falls back explicitly instead of pretending it was used. The Front-end Visual Director

🚀 How to Use

  1. Installation: Copy the desired skill folder (e.g., HighTriad) into your Agent's skills/ directory.
  2. Activation: Add the SKILL.md content of the chosen skill to your Agent's System Prompt or Context.
  3. Usage: The agent adopts the persona and utilizes the linked tools/references automatically.

📂 Repository Structure

.
├── Dialectical_Materialism/ # [Skill] Philosophical Reasoning
│ ├── SKILL.md
│ └── references/ # The Three Laws of Dialectics
├── Meihua_Yishu/ # [Skill] Hybrid AI Divination
│ ├── SKILL.md
│ ├── scripts/ # Calculation & Gemini Integration
│ └── references/ # Hexagrams & Strategy Database
├── AI-search-browser-use-main/ # [Skill] AI-assisted browser research
│ ├── SKILL.md
│ ├── ai_query.py
│ ├── scripts/ # Browser planning and page cleanup helpers
│ └── references/ # AI search target notes
├── HighTriad/ # [Skill] Three-High Architecture (Concurrency/Performance/Availability)
│ ├── SKILL.md
│ └── references/ # Tech/Industry/Templates
├── story-commentary-workflow/ # [Skill] Story commentary planning
│ ├── SKILL.md
│ └── references/ # Chapter schema, templates, and playbook
├── story-worldbuilding-planner/ # [Skill] Story and worldbuilding planner
│ └── SKILL.md
├── European-Chinese-Cleaner/ # [Skill] Translationese-to-native-Chinese rewrite
│ └── SKILL.md
├── PUA/ # [Skill] AI-Agent Under-Pressure Compliance Mode
│ └── SKILL.md
├── agent-fault-retrospective/ # [Skill] Agent fault retrospective and instruction-file guardrails
│ └── SKILL.md
├── Impeccable-design-ui/ # [Skill] Impeccable Design Workflow Hub
│ └── SKILL.md
├── awesome-design-setup/ # [Skill] Brand Design Picker (58 brands)
│ ├── SKILL.md
│ ├── references/ # Brand catalog with path keys
│ └── scripts/ # fetch_design.sh download helper
├── frontend-imagegen-director/ # [Skill] Generate front-end visuals from project style and build the page from the approved image
│ └── SKILL.md
└── README.md # Project Overview

🛡️ License

MIT License. Free to use for both personal and commercial Agent development.


Powered by Antigravity-Team & K-Dense Inc.

Releases

No releases published

Packages

 
 
 

Contributors