hoomanjs 1.26.2 → 1.26.3

package/README.md

@@ -418,6 +418,8 @@ OpenAI-compatible gateways that put token `usage` on the last streamed chunk tog
 
 Use for a [TensorZero](https://tensorzero.com) gateway’s **OpenAI-compatible** HTTP API (for example `/openai/v1`). Same general `params` shape as **OpenAI** (`apiKey`, `clientConfig.baseURL`, etc.), but implemented with **StrandsTensorZeroModel** so gateway-specific streaming works end-to-end: reasoning carried in `tensorzero_extra_content` shows as thinking in the chat UI, and usage metadata is normalized.
 
+Gateway features such as **tag-scoped rate limits** use TensorZero tags on each request. Set them in config under the nested `params` object (Strands forwards that object onto the Chat Completions body), for example `tensorzero::tags.user_id`.
+
 ```json
 {
   "provider": "tensorzero",
@@ -426,6 +428,11 @@ Use for a [TensorZero](https://tensorzero.com) gateway’s **OpenAI-compatible**
     "apiKey": "your-tensorzero-or-gateway-key",
     "clientConfig": {
       "baseURL": "http://localhost:3000/openai/v1"
+    },
+    "params": {
+      "tensorzero::tags": {
+        "user_id": "your-stable-user-or-tenant-id"
+      }
     }
   }
 }

package/dist/core/models/tensorzero/index.d.ts

@@ -5,5 +5,8 @@ export type TensorZeroModelConfig = Omit<OpenAIModelOptions, "api" | "modelId">;
 /**
  * Strands {@link Model} for a TensorZero gateway’s OpenAI-compatible HTTP API (`/openai/v1`).
  * Handles gateway-specific streaming: token usage on the last chunk and `tensorzero_extra_content` thoughts.
+ *
+ * Set TensorZero tags (e.g. `user_id` for gateway rate limits) in config under `llm.params.params`:
+ * `"params": { "tensorzero::tags": { "user_id": "…" } }` alongside `apiKey` / `clientConfig`.
  */
 export declare function create(model: string, params: Record<string, unknown>): StrandsTensorZeroModel;

package/dist/core/models/tensorzero/index.js

@@ -3,6 +3,9 @@ export { StrandsTensorZeroModel } from "./strands-tensorzero.js";
 /**
  * Strands {@link Model} for a TensorZero gateway’s OpenAI-compatible HTTP API (`/openai/v1`).
  * Handles gateway-specific streaming: token usage on the last chunk and `tensorzero_extra_content` thoughts.
+ *
+ * Set TensorZero tags (e.g. `user_id` for gateway rate limits) in config under `llm.params.params`:
+ * `"params": { "tensorzero::tags": { "user_id": "…" } }` alongside `apiKey` / `clientConfig`.
  */
 export function create(model, params) {
     return new StrandsTensorZeroModel({

package/dist/core/models/tensorzero/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../../src/core/models/tensorzero/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAC;AAEjE,OAAO,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAC;AAIjE;;;GAGG;AACH,MAAM,UAAU,MAAM,CACpB,KAAa,EACb,MAA+B;IAE/B,OAAO,IAAI,sBAAsB,CAAC;QAChC,GAAG,EAAE,MAAM;QACX,OAAO,EAAE,KAAK;QACd,GAAI,MAAgC;KACrC,CAAC,CAAC;AACL,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../../src/core/models/tensorzero/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAC;AAEjE,OAAO,EAAE,sBAAsB,EAAE,MAAM,yBAAyB,CAAC;AAIjE;;;;;;GAMG;AACH,MAAM,UAAU,MAAM,CACpB,KAAa,EACb,MAA+B;IAE/B,OAAO,IAAI,sBAAsB,CAAC;QAChC,GAAG,EAAE,MAAM;QACX,OAAO,EAAE,KAAK;QACd,GAAI,MAAgC;KACrC,CAAC,CAAC;AACL,CAAC"}

package/dist/core/skills/built-in/hooman-config/SKILL.md

@@ -173,7 +173,7 @@ Provider notes (these refer to fields inside `options.params` unless noted):
 - `moonshot`: `params.apiKey`, `baseURL`, `headers`, and `fetch` configure the provider. Other keys are forwarded as Vercel model config.
 - `xai`: `params.apiKey`, `baseURL`, and `headers` configure the provider. Other keys are forwarded as Vercel model config.
 - `openai`: Strands **OpenAIModel** (Chat Completions). `params.apiKey` (or env `OPENAI_API_KEY`), optional `clientConfig` (e.g. `baseURL` for an OpenAI-compatible HTTP API). `model` becomes `modelId`. Streaming usage from some gateways is normalized automatically. For a **TensorZero** gateway, prefer **`tensorzero`** if you want reasoning/thinking streamed into the UI and consistent usage metadata.
-- `tensorzero`: Same Chat Completions shape as `openai`, but **`StrandsTensorZeroModel`** — use for a [TensorZero](https://tensorzero.com) gateway’s OpenAI-compatible route (e.g. `clientConfig.baseURL` like `http://localhost:3000/openai/v1`). Maps gateway `tensorzero_extra_content` thoughts to Strands reasoning deltas and normalizes final-chunk usage. `model` is typically `tensorzero::function_name::<name>` (or other TensorZero model id strings the gateway accepts).
+- `tensorzero`: Same Chat Completions shape as `openai`, but **`StrandsTensorZeroModel`** — use for a [TensorZero](https://tensorzero.com) gateway’s OpenAI-compatible route (e.g. `clientConfig.baseURL` like `http://localhost:3000/openai/v1`). Maps gateway `tensorzero_extra_content` thoughts to Strands reasoning deltas and normalizes final-chunk usage. `model` is typically `tensorzero::function_name::<name>` (or other TensorZero model id strings the gateway accepts). For gateway tag-based rate limits, set `params.params["tensorzero::tags"]` (e.g. `user_id`) in the model entry in config.
 - `ollama`: `params.host`, `keepAlive`, `options`, and `think` configure the Ollama wrapper. `think` may be `true`, `false`, `"high"`, `"medium"`, or `"low"`.
 - `bedrock`: `params.region`, `clientConfig`, and optional `apiKey` configure Bedrock access. Put AWS credentials under `params.clientConfig.credentials` with `accessKeyId`, `secretAccessKey`, and optional `sessionToken`; put an AWS CLI/shared-config profile name in `params.clientConfig.profile`. If credentials and profile are omitted, Bedrock uses the AWS SDK default credential chain, including environment variables and AWS CLI/shared credentials. Other keys are forwarded as Bedrock model options, such as `temperature`, `maxTokens`, `stream`, and `cacheConfig`.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "hoomanjs",
-  "version": "1.26.2",
+  "version": "1.26.3",
   "description": "Hackable AI agent toolkit for building local CLI, ACP, MCP, and channel-driven workflows.",
   "author": {
     "name": "Vaibhav Pandey",