Plori

Official

by plori-ai

Overview Schema Source Related Servers Discussions Admin

JavaScript

Remote

Steer a running worker or run

steer_workspace_worker

Destructive

Send a message into a running worker's turn to correct its direction without stopping it; the agent reads it before its next step and continues the same run.

Instructions

Send a message into a worker's running turn, for example to correct its direction, without stopping it. run_id is the worker's run_id from get_workspace_worker; a run started by send_workspace_message is also accepted. The agent reads the message before its next step and continues the same run; disposition steering means the running turn holds the message. A run that is not running (not started yet, paused on an approval or question, or ended) is refused with 409 run_not_running: answer a paused run with answer_pending_input. If the running turn cannot take the message, for example because it is ending, the message is withdrawn and the call is refused with 409 steer_not_delivered. A message sent through this connector makes the run's later outward actions pause for the account owner's approval, as for a run started through this connector. idempotency_key identifies the message: the same key with the same message returns the first result instead of sending it twice.

Input Schema

Name	Required	Description
`run_id`	Yes	The running run to send the message to.
`message`	Yes	The message for the agent, in plain text (at most 65536 bytes).
`idempotency_key`	No	Optional string of 1-255 bytes you generate for this message. Omitted, each call sends a new message.

Output Schema

Name	Required	Description
`run_id`	Yes	The run that holds the message.
`steer_id`	Yes	The id of the delivered message.
`disposition`	Yes	steering: the running turn holds the message and the agent reads it before its next step. applied: the agent has already read it (a repeated request with the same idempotency_key).

Schema Changelog

Changes observed during successful MCP inspections.

Addedv0.17.1Oct 18, 2026

TDQS

A4.8/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Adds a great deal beyond the annotations: two distinct 409 failure modes (run_not_running, steer_not_delivered), the withdrawal of the message on undeliverable turns, the approval-pause side effect on subsequent outward actions, and the fact that the agent reads the message at its next step. The conditional idempotency via idempotency_key is also disclosed, qualifying the annotation hint rather than contradicting it.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Purpose is front-loaded and every sentence carries operational weight with no filler. It is dense and delivered as one long block, so structural breaks around errors, side effects and idempotency would improve scanability, but nothing is wasted.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With an output schema present, return values need not be described, and the description fills the remaining gaps: failure codes, state preconditions, side effects and alternative routing. An agent has everything needed to call this correctly on the first attempt.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3, but the description still adds provenance for run_id (from get_workspace_worker, or a run started by send_workspace_message) and clarifies the conditional idempotency semantics of idempotency_key ('omitted, each call sends a new message').

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Send a message into a worker's running turn') plus the defining constraint 'without stopping it'. This cleanly separates it from send_workspace_message (starting a run), answer_pending_input (answering a paused run) and cancel_run (stopping).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly names the conditions under which it applies (a running turn) and when it is refused (not started, paused on approval/question, or ended) and routes the agent to the correct alternative, answer_pending_input, for the paused case. No reliance on inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.