- Opencode (cli) = cline -> openwork (ui) = claude cowork
|
repository |
version |
stars |
contributors |
install |
system |
inferencer |
Db (memory) |
其他 |
|
openclaw |
2026.3.24 (> 1) |
336k |
1316 |
|
WSL |
|
|
|
|
nanobot |
0.1.4.post5 |
36.3k |
177 |
不需 Docker |
Windows |
llama.cpp |
X |
學術背景 |
|
picoclaw |
0.2.4 |
26.1k |
188 |
|
|
|
|
|
|
zeroclaw |
0.6.2 |
28.8k |
204 |
不需 Docker |
Windows |
llama.cpp |
sqlite |
|
|
ironclaw |
0.21.0 - 2026-03-20 |
10.9k |
97 |
不需docker |
Windows/wsl/linux/macOS |
llama.cpp |
pgvector |
NEAR AI account |
|
nanoclaw |
1.2.0 |
25.5k |
67 |
docker |
WSL2 |
Anthropic Claude Agent SDK |
|
|
|
cline |
3.75.0 |
59.4k |
307 |
不需 Docker |
Windows |
llama.cpp |
|
VS Code 擴充功能 |
zeroclaw
https://www.53ai.com/news/Openclaw/2026030832065.html
https://github.com/zeroclaw-labs/zeroclaw/releases/tag/v0.6.2
- zeroclaw-x86_64-pc-windows-msvc.zip
Local LLMs for OpenClaw: a practical guide
openclaw 怎麼跟notion溝通
- Skill
- 指令取得notion api key
- 透過api存取notion資料
- 如果要讓這樣的agent存取向量資料庫,也是用skills ?
- 這樣的skills要具備什麼能力?
- 將問題轉成向量
- 將問題向量送到向量資料庫(by API)
- 接收回傳的文字
- 如果remote只有向量資料庫,Advanced RAG (混和檢索、重新排序等技術) 要在哪裡處理?
- 一個含有Adanved RAG模組的skill,而且要能跟現有架構相容
- Openclaw的advanced RAG是用在記憶檢索
- https://github.com/openclaw/openclaw/issues/3170 (Closed as not planned)
- https://github.com/openclaw/openclaw/issues/4270 (duplicated, Closed)
- https://laplusda.com/posts/openclaw-memory-upgrade-qmd/
- 加入 qmd 讓openclaw的記憶檢索使用qmd的重排序
- 這樣的skills要具備什麼能力?
- 沒看到針對vector database的skills
- Google: vector database skills
- https://mcpmarket.com/zh/tools/skills/vector-database-management-1
- Activate for Pinecone, Weaviate, Chroma, pgvector, RAG, and similarity search.
- https://github.com/VoltAgent/awesome-agent-skills
- https://github.com/sanjay3290/ai-skills/tree/main/skills/postgres
- Execute safe read-only SQL queries against PostgreSQL databases
- https://github.com/sanjay3290/ai-skills/tree/main/skills/postgres
- https://mcpmarket.com/zh/tools/skills/vector-database-management-1
- Google: vector database skills
Github issue學習
- 撰寫架構 (ex. https://github.com/openclaw/openclaw/issues/3170)
- 總結summary
- 解決方案proposed solution
- 程式碼
- 考慮的替代方案Alternatives considered
- 補充訊息Additional context
- 證據Evidence
- 向後相容性Backward compatibility
- 如何全面了解並清楚的解釋現存解決方案的優缺點,而做出選擇
- Deep Agents: The Harness Behind Claude Code, Codex, Manus, and OpenClaw
- Deepagents 安裝
- Ironclaw安裝
- CAMEL-AI = deepagents ?
- Rust 生態與 SA Python 棧不相容 -> 修改prompt
-
- 模型以為SA的架構已經是確定的
Comments...
No Comments Yet...