@page-agent/llms 0.2.1 → 0.2.3

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.
Files changed (3) hide show
  1. package/LICENSE +2 -1
  2. package/package.json +1 -1
  3. package/README.md +0 -41
package/LICENSE CHANGED
@@ -1,6 +1,7 @@
1
1
  MIT License
2
2
 
3
- Copyright (c) 2025 Alibaba
3
+ Copyright (c) 2026 Alibaba
4
+ Copyright (c) 2026 Simon
4
5
 
5
6
  Permission is hereby granted, free of charge, to any person obtaining a copy
6
7
  of this software and associated documentation files (the "Software"), to deal
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@page-agent/llms",
3
- "version": "0.2.1",
3
+ "version": "0.2.3",
4
4
  "type": "module",
5
5
  "main": "./dist/lib/page-agent-llms.js",
6
6
  "module": "./dist/lib/page-agent-llms.js",
package/README.md DELETED
@@ -1,41 +0,0 @@
1
- # @page-agent/llms
2
-
3
- LLM client with a **reflection-before-action** mental model for page-agent.
4
-
5
- ## Why This Package Exists
6
-
7
- The LLM module and the agent logic are inherently coupled. This package exists not to decouple them, but to **define the interface contract** between the LLM and the agent.
8
-
9
- The core abstraction is the `MacroToolInput` — a structured output format that **forces the model to reflect before acting**.
10
-
11
- ## The Reflection-Before-Action Model
12
-
13
- Every tool call must first output its reasoning state before the actual action:
14
-
15
- ```typescript
16
- interface MacroToolInput {
17
- // Reflection (mandatory before any action)
18
- evaluation_previous_goal?: string // How well did the previous action work?
19
- memory?: string // Key information to remember
20
- next_goal?: string // What to accomplish next
21
-
22
- // Action (the actual operation)
23
- action: Record<string, any>
24
- }
25
- ```
26
-
27
- This design ensures that:
28
-
29
- 1. **The model evaluates its previous action** before deciding the next step
30
- 2. **Working memory is explicitly maintained** across conversation turns
31
- 3. **Goals are clearly stated**, making the agent's reasoning transparent and debuggable
32
-
33
- ## Key Components
34
-
35
- | Export | Description |
36
- |--------|-------------|
37
- | `LLM` | Main LLM client class with retry logic |
38
- | `MacroToolInput` | The reflection-before-action input schema |
39
- | `AgentBrain` | Agent's thinking state (eval, memory, goal) |
40
- | `LLMConfig` | Configuration for LLM connection |
41
- | `parseLLMConfig` | Parse and apply defaults to config |