@zhin.js/service-typesafe 1.0.5 → 1.0.6

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 (2) hide show
  1. package/README.md +12 -12
  2. package/package.json +5 -5
package/README.md CHANGED
@@ -24,25 +24,20 @@ zhin doctor
24
24
 
25
25
  The wizard keeps your chat model, writes the key to `.env`, installs the plugin manifest entry and starts Skill / Tool recommendations in `shadow` mode. On first setup, Agent selection, memory reranking and automatic approval start `off`. Reconfiguration preserves existing task settings and thresholds. Separate instances use separate key environment variables; an existing explicit key reference is preserved.
26
26
 
27
- For manual configuration, declare the plugin in `package.json`:
27
+ For manual configuration, declare the plugin in `zhin.config.yml`:
28
28
 
29
- ```json
30
- {
31
- "zhin": {
32
- "plugins": [{ "package": "@zhin.js/service-typesafe", "instanceKey": "typesafe" }]
33
- }
34
- }
29
+ ```yaml
30
+ configVersion: 2
31
+ plugins:
32
+ typesafe: '@zhin.js/service-typesafe'
35
33
  ```
36
34
 
37
35
  Add its instance settings and exact binding in `zhin.config.yml`, alongside the existing AI configuration:
38
36
 
39
37
  ```yaml
38
+ configVersion: 2
40
39
  plugins:
41
- typesafe:
42
- apiKey: ${TYPESAFE_API_KEY}
43
- model: jev-latest
44
- timeoutMs: 10000
45
- maxRetries: 2
40
+ typesafe: "@zhin.js/service-typesafe"
46
41
  ai:
47
42
  # Keep existing agents/providers and other AI settings.
48
43
  decisions:
@@ -51,6 +46,11 @@ ai:
51
46
  mode: shadow
52
47
  tools:
53
48
  mode: shadow
49
+ typesafe:
50
+ apiKey: ${TYPESAFE_API_KEY}
51
+ model: jev-latest
52
+ timeoutMs: 10000
53
+ maxRetries: 2
54
54
  ```
55
55
 
56
56
  The instance registers its client through `provideAgentDecisionProvider` from `@zhin.js/agent/runtime` in its own plugin scope. The registration binds the resource to its owning instance, so an inherited ancestor provider cannot satisfy a different configured instance. Agent configuration must explicitly bind the instance through the runtime decision configuration; merely installing a provider does not authorize decisions or change existing routing. Independent instances have independent clients and lifecycle cleanup.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@zhin.js/service-typesafe",
3
- "version": "1.0.5",
3
+ "version": "1.0.6",
4
4
  "description": "Optional TypeSafe AI Jev decision provider for Zhin",
5
5
  "type": "module",
6
6
  "main": "./lib/index.js",
@@ -24,15 +24,15 @@
24
24
  "@zhin.js/ai": "1.1.41"
25
25
  },
26
26
  "peerDependencies": {
27
- "@zhin.js/agent": "^1.1.35",
28
- "zhin.js": "^1.1.10"
27
+ "@zhin.js/agent": "^1.1.36",
28
+ "zhin.js": "^1.1.11"
29
29
  },
30
30
  "devDependencies": {
31
31
  "@types/node": "^26.1.2",
32
32
  "typescript": "^6.0.3",
33
33
  "vitest": "^4.1.10",
34
- "zhin.js": "1.1.10",
35
- "@zhin.js/agent": "1.1.35"
34
+ "@zhin.js/agent": "1.1.36",
35
+ "zhin.js": "1.1.11"
36
36
  },
37
37
  "engines": {
38
38
  "node": ">=22.19.0"