hoomanjs 1.12.0 → 1.12.1
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 +25 -17
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
<div align="center">
|
|
2
2
|
<h1>Hooman</h1>
|
|
3
3
|
<p>
|
|
4
|
-
Hooman is a Bun-powered
|
|
4
|
+
Hooman is a hackable, Bun-powered AI agent toolkit for local workflows. It is built with TypeScript, <a href="https://www.npmjs.com/package/@strands-agents/sdk">Strands Agents SDK</a>, and <a href="https://github.com/vadimdemedes/ink">Ink</a>.
|
|
5
5
|
</p>
|
|
6
6
|
<p>
|
|
7
7
|
<a href="https://bun.com"><img src="https://img.shields.io/badge/runtime-Bun-f9f1e1?logo=bun&logoColor=000000" alt="Bun" /></a>
|
|
@@ -16,12 +16,12 @@
|
|
|
16
16
|
</p>
|
|
17
17
|
</div>
|
|
18
18
|
|
|
19
|
-
It gives you:
|
|
19
|
+
It gives you a practical toolkit to build and run agent workflows:
|
|
20
20
|
|
|
21
21
|
- a one-shot `exec` command for single prompts
|
|
22
|
-
- a stateful `chat` interface for
|
|
23
|
-
- a `daemon` command for
|
|
24
|
-
- an Ink-powered `configure` workflow for
|
|
22
|
+
- a stateful `chat` interface for iterative sessions
|
|
23
|
+
- a `daemon` command for channel-driven MCP automation
|
|
24
|
+
- an Ink-powered `configure` workflow for app config, prompts, MCP servers, and installed skills
|
|
25
25
|
- an `acp` command for running Hooman as an Agent Client Protocol (ACP) agent over stdio
|
|
26
26
|
|
|
27
27
|
## Features
|
|
@@ -32,6 +32,7 @@ It gives you:
|
|
|
32
32
|
- MCP server `instructions` support: server-provided instructions are appended to the agent system prompt
|
|
33
33
|
- MCP channel notification support through `hooman daemon --channels`
|
|
34
34
|
- Skill discovery / install / removal through the integrated configure flow
|
|
35
|
+
- Toolkit-oriented architecture with configurable tools, prompts, memory, and transports
|
|
35
36
|
- Interactive terminal UI for chat and configuration
|
|
36
37
|
|
|
37
38
|
## Requirements
|
|
@@ -157,18 +158,20 @@ hooman daemon --channels --yolo
|
|
|
157
158
|
|
|
158
159
|
### Feature Flags
|
|
159
160
|
|
|
160
|
-
Runtime tools and prompt sections are controlled from `config.json` under `
|
|
161
|
+
Runtime tools and prompt sections are controlled from `config.json` under `tools`:
|
|
161
162
|
|
|
162
|
-
- `
|
|
163
|
-
- `
|
|
164
|
-
- `
|
|
165
|
-
- `
|
|
166
|
-
- `
|
|
163
|
+
- `tools.fetch.enabled`
|
|
164
|
+
- `tools.filesystem.enabled`
|
|
165
|
+
- `tools.shell.enabled`
|
|
166
|
+
- `tools.ltm.enabled`
|
|
167
|
+
- `tools.wiki.enabled`
|
|
168
|
+
- `tools.mcp.enabled` (enables MCP management tools + prefixed MCP server tools/instructions)
|
|
169
|
+
- `tools.skills.enabled` (enables skills management tools + skills prompt sections)
|
|
167
170
|
|
|
168
171
|
Both `ltm` and `wiki` include dedicated Chroma settings under:
|
|
169
172
|
|
|
170
|
-
- `
|
|
171
|
-
- `
|
|
173
|
+
- `tools.ltm.chroma` (default collection: `memory`)
|
|
174
|
+
- `tools.wiki.chroma` (default collection: `wiki`)
|
|
172
175
|
|
|
173
176
|
### `hooman configure`
|
|
174
177
|
|
|
@@ -211,7 +214,7 @@ Hooman stores its data in:
|
|
|
211
214
|
|
|
212
215
|
Important files and folders:
|
|
213
216
|
|
|
214
|
-
- `config.json` - app name, LLM provider/model, tool
|
|
217
|
+
- `config.json` - app name, LLM provider/model, tool flags, LTM/wiki settings, compaction
|
|
215
218
|
- `instructions.md` - system instructions used to build the agent prompt
|
|
216
219
|
- `mcp.json` - MCP server definitions
|
|
217
220
|
- `skills/` - installed skills
|
|
@@ -231,9 +234,6 @@ This is the shape managed by `hooman configure`:
|
|
|
231
234
|
"params": {}
|
|
232
235
|
},
|
|
233
236
|
"tools": {
|
|
234
|
-
"allowed": []
|
|
235
|
-
},
|
|
236
|
-
"features": {
|
|
237
237
|
"fetch": {
|
|
238
238
|
"enabled": true
|
|
239
239
|
},
|
|
@@ -260,6 +260,12 @@ This is the shape managed by `hooman configure`:
|
|
|
260
260
|
"wiki": "wiki"
|
|
261
261
|
}
|
|
262
262
|
}
|
|
263
|
+
},
|
|
264
|
+
"mcp": {
|
|
265
|
+
"enabled": false
|
|
266
|
+
},
|
|
267
|
+
"skills": {
|
|
268
|
+
"enabled": false
|
|
263
269
|
}
|
|
264
270
|
},
|
|
265
271
|
"compaction": {
|
|
@@ -269,6 +275,8 @@ This is the shape managed by `hooman configure`:
|
|
|
269
275
|
}
|
|
270
276
|
```
|
|
271
277
|
|
|
278
|
+
Tool approvals are session-scoped and are not persisted in `config.json`.
|
|
279
|
+
|
|
272
280
|
Supported `llm.provider` values:
|
|
273
281
|
|
|
274
282
|
- `ollama`
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "hoomanjs",
|
|
3
|
-
"version": "1.12.
|
|
4
|
-
"description": "Bun-powered
|
|
3
|
+
"version": "1.12.1",
|
|
4
|
+
"description": "Hackable Bun-powered AI agent toolkit for building local CLI, ACP, MCP, and channel-driven workflows.",
|
|
5
5
|
"author": {
|
|
6
6
|
"name": "Vaibhav Pandey",
|
|
7
7
|
"email": "contact@vaibhavpandey.com"
|