Skip to content

Commit f28ac1a

Browse files
authored
cagent update builtins (#24053)
- **cagent: "script_shell" toolset should be "script"** - **cagent: add user_prompt toolset** --------- Signed-off-by: David Karlsson <35727626+dvdksn@users.noreply.github.com>
2 parents 120b064 + 63e1e61 commit f28ac1a

1 file changed

Lines changed: 33 additions & 6 deletions

File tree

content/manuals/ai/cagent/reference/toolsets.md

Lines changed: 33 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -43,7 +43,7 @@ MCP toolsets
4343
(stdio) or remote servers (HTTP/SSE). MCP enables access to a wide ecosystem
4444
of standardized tools.
4545
Custom toolsets
46-
: Shell scripts wrapped as tools with typed parameters (`script_shell`). This
46+
: Shell scripts wrapped as tools with typed parameters (`script`). This
4747
lets you define domain-specific tools for your use case.
4848

4949
## Configuration
@@ -77,7 +77,7 @@ agents:
7777
Authorization: Bearer ${API_TOKEN}
7878

7979
# Custom shell tools
80-
- type: script_shell
80+
- type: script
8181
tools:
8282
build:
8383
cmd: npm run build
@@ -313,10 +313,37 @@ toolsets:
313313
- type: fetch
314314
```
315315

316+
### User Prompt
317+
318+
The `user_prompt` toolset lets your agent ask you questions during task
319+
execution. When the agent needs clarification, decisions, or additional
320+
information it can't determine on its own, it displays a dialog and waits for
321+
your response.
322+
323+
You'll see a prompt with the agent's question. Depending on what the agent
324+
needs, you might provide free-form text, select from options, or fill out a
325+
form with multiple fields. You can accept and provide the information, decline
326+
to answer, or cancel the operation entirely.
327+
328+
#### Configuration
329+
330+
```yaml
331+
toolsets:
332+
- type: user_prompt
333+
```
334+
335+
No additional configuration is required. The tool becomes available to the
336+
agent once configured. When the agent calls this tool, the user sees a dialog
337+
with the prompt. The user can:
338+
339+
- **Accept**: Provide the requested information
340+
- **Decline**: Refuse to provide the information
341+
- **Cancel**: Cancel the operation
342+
316343
### API
317344

318345
The `api` toolset lets you define custom tools that call HTTP APIs. Similar to
319-
`script_shell` but for web services, this allows you to expose REST APIs,
346+
`script` but for web services, this allows you to expose REST APIs,
320347
webhooks, or any HTTP endpoint as a tool your agent can use. The agent sees
321348
these as typed tools with automatic parameter validation.
322349

@@ -367,9 +394,9 @@ requests, parameters are sent as JSON in the request body.
367394

368395
Supported argument types: `string`, `number`, `boolean`, `array`, `object`.
369396

370-
### Script Shell
397+
### Script
371398

372-
The `script_shell` toolset lets you define custom tools by wrapping shell
399+
The `script` toolset lets you define custom tools by wrapping shell
373400
commands with typed parameters. This allows you to expose domain-specific
374401
operations to your agent as first-class tools. The agent sees these custom
375402
tools just like built-in tools, with parameter validation and type checking
@@ -385,7 +412,7 @@ parameters:
385412

386413
```yaml
387414
toolsets:
388-
- type: script_shell
415+
- type: script
389416
tools:
390417
deploy:
391418
cmd: ./deploy.sh

0 commit comments

Comments
 (0)