YunyueLi/SecondU
A personal agent built around a continuously maintained digital twin. Local context, specialist teams and lasting work.
笑えばいいと思うよ。
A personal agent built around a continuously maintained digital twin. Local context, specialist teams and lasting work.
Seven agent skills for interview prep, a 27-role knowledge base, and an open Markdown workspace format. 面试准备技能包 + 岗位知识库 + 开放工作台格式。
modelese(模型腔)— lexicon, linter & rewrite-gate for the private dialect AI develops on long runs. 新时代文言文检测器:给 AI 黑话打分、打回、翻成人话。CC0 词典。
Knowledge graphs from fiction — every edge quotes the page it came from. 从虚构文本生成知识图谱:每条边都能溯源到原文。
Design Teardowns — 21 landing pages taken apart into verifiable design evidence: extracted tokens, assets, shaders and timelines, measured to the last value.
The setup I use for pi. Changes a ton, don't recommend hard copying it, just a taste of what's possible with it
作家蒸馏 Voice Capsules — 把文学巨匠蒸馏成可调用、可评测的「声部封装」:七层模型+负空间+判别器自举闭环 / Distilling literary masters into callable, evaluable voice capsules
A transparent, tier-weighted Honor Index that ends the GOAT debate with numbers. 竞技荣誉的可视化殿堂:透明可审计的荣誉指数,用数据终结 GOAT 之争(首发英雄联盟)。
One command installs a curated set of Claude Code design skills — plus a decision table for which skill to use when.
Claude Design DESIGN.md prompts by aesthetic family, remix recipes, skills, video teardowns, X signal, honest community takes.
Offline, auditable, fail-closed macOS WeChat external storage manager
Show usage stats for OpenAI Codex and Claude Code, without having to login.
Agent skill for beautiful, verifiable architecture, workflow, sequence, data-flow, and lifecycle diagrams—self-contained HTML with motion and crisp export.
Unofficial best-effort repair tool for local Codex conversation history after account or provider changes.
Profile README
Template for reproducible research-paper projects: src/data/figures/paper structure, provenance disclosure, LaTeX skeleton
Telos Community Edition: reverse-design learning engine and runnable local-first product
The open-source AI growth team for user acquisition — every decision rule-based and auditable. 开源的 AI 用户增长团队:决策由规则计算、可审计,方向盘在你手里。
Homebrew tap for ClaudeRoam.
Sandbox repo for experiments
中文小说/文案去「AI腔」自查:grep揪套路的清单 + 贴任意AI的prompt + Claude Code skill。A battle-tested Chinese-prose de-AI-slop checklist + Claude Code skill.
A Simple and Universal Swarm Intelligence Engine, Predicting Anything. 简洁通用的群体智能引擎,预测万物
微舆:人人可用的多Agent舆情分析助手,打破信息茧房,还原舆情原貌,预测未来走向,辅助决策!从0实现,不依赖任何框架。