Research & Data

A staged, evidence-driven research-agent skills framework for academic and scientific workflows.

AI 产出验收清单生成器（四维度查错）：按功能/接口/安全/回归出可勾选验收清单。适合 code review、AI 外包验收。

Local-first watchdog and Codex skill for stalled tool calls, long-running task monitoring, oversized-thread diagnosis, bounded cleanup, and safe handoff recovery.

字帖/作业纸生成器（米字格田字格13种）：一键生成可打印中文字帖与练习纸，支持汉字居中、拼音标注。适合书法练字、小学语文作业、早教字卡。

中英文字段名对齐映射工具：把杂乱字段名（中/英/驼峰/下划线）映射到规范 key，支持别名与模糊匹配。适合对接第三方接口、ERP 字段整理、数据清洗。

name: apm

RPA 拟人化鼠标轨迹/打字（防机器人识别）：贝塞尔轨迹、随机延时、犹豫点击、逐字打字。适合 RPA 测试、批量操作防封。

Claude skill: builds a live GA4 + Search Console + PageSpeed Insights analytics dashboard for any web project

Reusable Claude skills to help students study more effectively.

Claude skills for evidence-based career self-analysis — blindspots, goal alignment, level gaps. Local & private. Your data. Your eyes only.

A collection of reusable Claude Skills for product growth and commercial strategy — including a parameterized competitor intelligence system (weekly research, threat scoring, and role-based leadership reporting). Built for practical use, not demos.

Custom skills for Claude Code — fable-deepresearch (2-pass verified deep research) + knowledge-cite/absorb/query

A cross-agent marketplace for plugins maintained in this repository. It supports OpenAI Codex, Claude Code, and Pi while keeping each agent's native package metadata alongside shared Agent Skills.

A project-agnostic Python framework for evaluating, benchmarking, and optimizing AI agent skills.

Codex skill for running, resuming, inspecting, and reporting PyTorch pytest workflows

Recently updated agent-skill-related GitHub repository: ImagingDataCommons/imaging-data-commons-skill.

A股 + 港股专业财经研究 Skill：基本面、财报、估值、排雷、产业链与打新，生成有来源、可证伪的研究报告。

A Claude skill that turns your study notes into high-quality, cloze-only Anki flashcards; Evidence-based spaced repetition, delivered as an easily importable TSV.

Codex skill for publication-grade Python MRI machine learning with leakage-safe SHAP analysis

Multi-agent delivery workflow for Codex and SKILL.md-compatible coding agents

GitHub repository health reports in your terminal and AI agent — health scores, deep scans, fake-star audits from trustoss.org

My personal Claude Skills, unified in one repo — battle-tested across work and life, and exhibited on Anvil.

AI-assisted Windows disk auditing and conservative cleanup skill for Claude Code, Codex, and other coding agents.

A disciplined engineering team for Claude Code — not a prompt. 11 specialist agents auto-chain plan → build → audit → ship, with security/privacy gates and approval-guarded commits. Set up a fresh project or safely adopt an existing repo.