Writing & Editing

A standalone CLI tool for n8n

Zero-effort Generative Engine Optimization toolkit. Rank higher in ChatGPT, Perplexity, Claude & Gemini with one command

Engineering Foundry Skill for structured AI-assisted engineering workflows

利用 `~/.codex/auth.json` 中的本地 Codex 凭证查询 ChatGPT/Codex 的速率限制重置额度（rate-limit reset credits）。此操作适用于用户查询速率限制重置额度、Codex 使用重置额度或访问 `chatgpt.com` 后端 API 的 `wham/rate-limit-reset-credits` 接口的场景；特别是在无需输出令牌（token）、Cookie 或完整唯一 ID，且希望将 UTC 时间戳转换为北京时间的情况下使用。

Codex skill for online experiment Results analysis and writing

Codex skill for building editable paper figures with local Microsoft Visio on Windows.

Custom Claude skills for professional writing and job applications — personal voice, CV/cover letter methodology, and AI writing detection.

🔬 A 13-agent AI research team for Claude — clarifies your question, searches the web/docs/Drive, verifies every source, red-teams its conclusions, and writes a cited report. Multi-agent deep research as a Claude Skill.

Codex Skill for converting Axure/HTML prototypes into editable Ardot files / 将 Axure/HTML 原型转换为可编辑 Ardot 文件的 Codex Skill

Install any MCP server into Claude Desktop without errors. Handles environment diagnosis, all auth types, config writing with absolute paths, and end-to-end verification. Includes references for 20 mainstream MCPs and a generic flow for any unlisted MCP.

A lightweight Codex skill that minimizes context-window usage by persisting project state to structured files (conductor/ + daily logs). It provides an idempotent setup script, concise progress/decision templates, and checkpoint workflows so teams can resume work across sessions with minimal chat history.

Claude skill: build a self-contained interactive HTML artifact (report/deck/flowchart/prototype/editor UI), save to ~/.claude/artifacts, pop it open. Bundles 20 Apache-2.0 examples from ThariqS/html-effectiveness.

🛡️ Insurance Underwriting & Claims Analyzer 📊 Evaluate insurance risk, underwriting decisions, and claim validity using actuarial science, regulatory standards, and evidence-based risk assessment.

record-label-manager

Package and publish local agent skills as open-source GitHub repositories

Offline CLI and agent skill to search, browse, and summarize local Claude Code and Codex session transcripts.

Humanize: an AI-text humanizer agent skill. 29-pattern detection, adversarial refinement, and a guaranteed zero em dashes.

Abre los endpoints de tu app sin ser experto: detecta tu stack, genera un endpoint REST de solo lectura y arma la propuesta para registrar tu fuente en la red civica de personas desaparecidas. CLI sin dependencias; funciona igual con Claude, Codex y Cursor.

Skill that writes judge-ready hackathon READMEs, grounded in your repo and mapped to the judging rubric.

Agent-assisted maintainership and development framework for Apache projects — Triage and Drafting (agent-authored fixes with human review); Mentoring, Pairing (developer-side dev-cycle), and Auto-merge on the roadmap.

Continual agent skill evolution through persistent decision history. Whole-skill optimisation (SKILL.md + scripts + references) with every decision landing as a local Git issue / PR / wiki. Runs on any agentskills.io runtime — Claude Code, Codex, OpenClaw, Hermes.

把投研逻辑发布成 Serverless 数据接口和可分享、会自动更新的量化看板。

Answer ultra-precise questions — and generate grounded, citation-checked reference docs — about any open-source project from its real source code, issues, PRs, docs and the web. Grounded retrieval, not the model's memory. A skills.sh agent skill + zero-dependency CLI.

Transform rough natural-language requests into clear, executable prompts for agent delegation, then dispatch them through Codex, Claude Code, or compatible host tools when available after confirmation unless direct sending is requested. Use when the user explicitly invokes $your-prompt-engineer, or naturally asks to write, improve, review, translate, structure, or prepare a prompt for an agent/task/subagent/worker; says "帮我写个 prompt", "整理成 agent prompt", "优化这个 prompt 再发给 agent", "write a prompt for an agent", "turn this into an agent prompt", or "agent に投げるプロンプトを書いて". If the user only asks to run/delegate/inspect/fix with an agent but does not mention prompt preparation, do not trigger implicitly. If the prompt target says project/current project and the workspace is empty or projectless, ask for the project path before preparing or dispatching.