Send a workspace action request to an agent, as a page would
Send a workspace action request into your conversation with an agent.
This is the request a workspace HTML page makes through window.mutiro.request: a named action with a JSON payload, attributed to a source page in the agent's workspace.
The agent handles it with respond_to_action.
Requests and responses stay out of the chat timeline but remain in the conversation history.
The payload is a JSON object given inline, as @file, or on stdin with "-". It defaults to {}.
With --wait the command holds the conversation stream open and prints the agent's response (result or error) once it arrives.
Examples:
Load a learning guide's state and wait for the result
mutiro user message action amelia learning_guide.load '{"guideId":"paxos-v1"}'
--source learning/paxos-v1/index.html --wait --json
Send a browser error report captured by a page's diagnostics panel
mutiro user message action amelia preview.report_errors @report.json
--source learning/paxos-v1/index.html --wait
mutiro user message action <agent> <action> [payload-json] [flags]
Load a learning guide's state and wait for the result
Send a browser error report captured by a page's diagnostics panel
| Flag | Type | Description |
|---|---|---|
| -h, --help | bool | help for action |
| --json | bool | Output result as JSON |
| --root | string | Workspace root of the source page: owner or user:<username> (default "owner") |
| --section | string | Workspace section of the source page (default "other") |
| --source | string | Workspace path of the page making the request, e.g. learning/guide/index.html (required) |
| --wait | bool | Hold the stream open and print the agent's response |
| --wait-timeout int | bool | Seconds to wait for the response before failing (default 180) |
| Flag | Type | Description |
|---|---|---|
| --config | string | config file (default is ./.mutiro.yaml in current directory) |
| --debug | bool | debug mode |
| -v, --verbose | bool | verbose output |