📊 At a Glance
📫 Connect
🤖 AI Toolchain
🎯 Manifesto
"I ship small, traceable fixes. The kind that close issues — not the kind that churn the codebase."
| NOW | Active in 8 OSS projects · ~120 PRs in May 2026 |
| BUILDING | Security & i18n hardening for AI desktop tools |
| NEXT | Deeper into MCP tooling & terminal-side agents |
Auto-aggregated from every PR I've ever opened — refreshed daily.
| Project | ⭐ | PRs | ✅ Merged | 🚀 Open |
|---|---|---|---|---|
multica-ai/multica |
33002 | 38 | 8 | 27 |
tinyhumansai/openhuman |
27716 | 30 | 25 | 2 |
code-yeongyu/oh-my-openagent |
59419 | 20 | 2 | 18 |
slopus/happy |
21204 | 20 | 0 | 20 |
obra/superpowers |
206145 | 19 | 0 | 2 |
rtk-ai/rtk |
53963 | 18 | 1 | 16 |
thedotmack/claude-mem |
78066 | 14 | 0 | 14 |
nexu-io/open-design |
51951 | 14 | 9 | 5 |
Tencent/TencentDB-Agent-Memory |
4135 | 13 | 2 | 8 |
anomalyco/opencode |
165167 | 5 | 0 | 5 |
multica-ai/andrej-karpathy-skills |
154474 | 1 | 0 | 1 |
code-yeongyu/lsp-tools-mcp |
1 | 1 | 0 | 1 |
| Focus | Pattern |
|---|---|
| 🇨🇳 Chinese-community pain points | zh-CN i18n · WeChat / Feishu / DingTalk integrations |
| 🐛 Traceable real bugs | Sentry / issue-backed regressions, never speculative refactors |
| 🛡 Security hardening | CORS, symlink escape, credential exposure |
- 🔧 Opened
anomalyco/opencode— fix(opencode): fail task on empty subagent output - 🔧 Opened
anomalyco/opencode— fix(opencode): expose config parse errors over http - 🔧 Opened
anomalyco/opencode— fix(opencode): stabilize git diff patch prefixes - 🔧 Opened
anomalyco/opencode— fix(opencode): wait for shell output drain - 🔧 Opened
anomalyco/opencode— fix(config): allow web permission object rules
Languages
Ecosystems
AI Tooling
🎯 个人宣言
"小修复,可追溯。专修 close issue 的那种,不写 churn 代码。"
- 活跃于 8 个开源项目,5 月已交付 ~120 个 PR
- 聚焦 AI 桌面工具的安全加固 与 中文本地化
- 计划深入 MCP 服务工具链 与 终端侧 agent
| 方向 | 模式 |
|---|---|
| 🇨🇳 中文社区痛点 | zh-CN i18n · 微信 / 飞书 / 钉钉集成 |
| 🐛 可追溯的真实 bug | 锚定 Sentry / Issue 的回归,不做投机重构 |
| 🛡 安全加固 | CORS · symlink escape · 凭据暴露 |
| 项目 | 简介 |
|---|---|
tinyhumansai/openhuman ⭐13k |
Tauri + Rust 个人 AI 桌面 |
multica-ai/multica |
开源 Managed Agents 平台 |
obra/superpowers |
Agentic Skills 框架 |
nexu-io/open-design |
Claude Design 开源替代 |
rtk-ai/rtk |
LLM token 节省 CLI(Rust) |
Tencent/TencentDB-Agent-Memory |
腾讯 AI Agent 长期记忆 |
slopus/happy |
Claude Code 手机端 · ACP 后端加固 |
thedotmack/claude-mem |
Claude Memory 工具 · Windows 跨平台修复 |
⚙ This README is auto-updated by GitHub Actions · always open to OSS chat

