Writing & Editing

A high-fidelity, local-first MCP server and protocol for storing, versioning, syncing, and evaluating AI agent prompts and skills

A curated list of AI-powered tools, MCP servers, and agent skills for scientific research. From literature search to manuscript submission.

Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads before you install.

11 个面向中文科研写作的 Claude Code 纪律 Skill。

Personal collection of Claude skills - growing as I discover patterns and solve real-world problems

An automated AI research-paper writer based off Google's PaperOrchestra paper's implementation through a skills - benchmark + autoraters using any coding agent (Claude Code, Cursor, Antigravity, Cline, Aider). No API keys, no LLM SDKs.

A PHP library for parsing and validating SKILL.md files or raw SKILL.md content.

🚧 BETA · Skills for the handoff → compact → resume workflow. Preserve key session context across compactions, with a confirmation gate on resume.

Exercise: Integrate Model Context Protocol with GitHub Copilot

Exercise: Integrate Model Context Protocol with GitHub Copilot

Exercise: Integrate Model Context Protocol with GitHub Copilot

Copy/paste detector for programming source code, supports 223 formats. AI-ready with token-efficient reporter, skill and MCP server.

Claude Code skill: convert files to Markdown with uvx markitdown

ClawShow MCP Server — AI-callable skills via Model Context Protocol. Hosted at mcp.clawshow.ai/sse

Claude Code skill: decide when to render output as HTML over Markdown — based on Thariq Shihipar's html-effectiveness examples. Bilingual (zh-CN / en).

Cold email deliverability scanner — 400+ spam trigger words, dual scoring (deliverability + ICP clarity), clean rewrite. Claude Code skill.

An AI-driven internship recommendation system that suggests relevant roles based on user skills and preferred location, with integrated SQLite-based history tracking to enable context-aware, personalized, and continuously improving recommendations.

Claude Code skill that keeps documentation and code in sync — auto-detects project type, enforces docs-first workflow, and audits coverage on every task.

Re-grow a SKILL.md loader + registry + matcher + linter from scratch in Go — one mechanism per chapter, each ending with an upstream content reading.

Docusaurus plugin for transforming SKILL.md files from multiple directories into MDX pages

bstack P20 — Cross-Model Adversarial Review Gate. The writer cannot judge the work. 3 strata: Codex cross-vendor / fresh subagent / composed adversarial-review skills. Anti-slop ≥7/10.

AI Agent 技能库 — 通用 Markdown 格式，任何 Agent 框架均可加载

Agent Skill that teaches Cursor & Claude Code to output single-file HTML for spatial, parallel, comparable, or interactive content. 9 categories, 20 patterns, 12 ready-to-use templates. Inspired by Thariq Shihipar's 'unreasonable effectiveness of HTML'.

This repository highlights my hands-on exploration of AI agent skills, OpenAI-based tools, and development workflows. I worked with VS Code, GitHub Copilot, Node.js, Python, OpenCode, Prophet, and Plotly, while studying concepts like RAG, vector databases, and agent automation through various docs, tutorials, and research.