Research & Data

Claude Code Agent Skill: diagnose & resolve Microsoft Defender (MsMpEng.exe) high CPU on Windows dev machines via IO-trigger correlation + safe exclusion generation

The open layer of agentbox.id — skills, protocols, and documents for a world where Agents act and converse on the network. Grounded in first principles: token economics, conversation rhythm, guardianship.

Xiaohongshu creator-account diagnostic reports from your own backend data — funnel health, benchmark posts, per-note fixes. Powered by webclaw3.

基于 Karpathy LLM Wiki 模式的智能体知识库运营 Skill：Ingest 摄入 / Query 回填 / Lint 六维体检，把外部知识蒸馏为智能体可调用的行为规则。

Agent-native shopping for extreme value: verifiable same-product price evidence, checkout, orders, delivery, and after-sales.

A local-first AI chat workspace for models, agents, skills, plugins, search, RAG, voice, memory, and artifacts.

Codex Skill for OZ weekly satisfaction reports

Codex skills for rigorous open mathematics research, research-program management, and Lean 4 formal verification (rigorous-open-math-research + manage-math-research-program + lean-verify plugin + math-research-workflow pipeline).

Plan defensible research stories, Results logic, and living experiment plans as an installable Codex skill

Bilingual Codex skill for integrated Meta, Google Search, GA4, and Shopify ecommerce analysis

Agent Skills for capturing and surfacing team decisions with Jupi (search/log/submit). Install once; auto-updates.

Four Claude skills built to remove recurring manual work from actual, repeated tasks in my coursework, FYP search, job applications, and project presentations at FAST-NUCES Lahore.

Lint, score, secure and benchmark AI Agent Skills and instruction files across Codex, Claude Code, Cursor, Gemini CLI and GitHub Copilot.

亚马逊选品 之 Listing全维度穿透分析报告 加上 全品类分析 ，关键词分析，差评分析 ，市场调研 等等。OpenClaw /claude code agent skill, amazon sellersprite research MCP 智能体skill.基于 卖家精灵 MCP 服务和 Claude Skills 的亚马逊竞品分析工具集

Versioned Codex skill for evidence-backed historic place vitality research, scoring, and Office deliverables.

Codex skill: authorized web security assessment and evidence-based reporting

Codex skill for researching, drafting, previewing, and safely publishing Cantonese market commentary on TTM

Claude and Codex skill to analyze and evaluate VMware RVTools XLSX exports.

A Claude skill that turns rough topics into evidence-backed, research-ready project ideas. Use it when you need a thesis/capstone/final-year topic, want to find research gaps, check whether an idea is already explored or patented, or turn a vague idea into a well-scoped research proposal.

waterline - offline, LLM-free, dependency-free differential reachability analyzer for Agent Skill bundles: computes the reachable closure under four resolver models and reports where a scanner's view diverges from an agent's (the audit-runtime gap + reference laundering)

A Codex Skill for rigorous research orchestration, project governance, prompt compilation, and reproducible handoffs.

Evidence-first fact-checking Agent Skills suite for reproducible source verification and answer comparison.

TOA询问法：人指导AI的逐题征询协议 / A question-by-question elicitation protocol for human-guided AI

An Overview for Agency Agent to view all the Agents and Search for them