@wudao-org/core 0.4.20 → 0.4.21
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +10 -0
- package/package.json +1 -1
package/README.md
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
# @wudao-org/core
|
|
2
|
+
|
|
3
|
+
The engine inside [wudao-harness](https://github.com/feiniaojin/wudao-harness) — the agent loop, tool interface, session persistence, context compaction, and subagent machinery, with no dependency on any model vendor.
|
|
4
|
+
|
|
5
|
+
If you're building an agent runtime on top of a battle-tested loop, this is that loop: typed state transitions, streaming, append-only sessions you can resume and fork from any point, and cheapest-first compaction that never breaks your provider's prefix cache.
|
|
6
|
+
|
|
7
|
+
- Any OpenAI-compatible provider works; reasoning models (DeepSeek / GLM / Qwen / Kimi / MiniMax) are first-class.
|
|
8
|
+
- MIT license.
|
|
9
|
+
|
|
10
|
+
Want the whole thing ready to use? `npm install -g wudao-harness`.
|