Crevio

ask_crevio

Destructive

Delegate a job to the Crevio agent and wait for the result. The agent works your account through the Crevio API and every connected integration (products, customers, orders, email, socials, sites, research) and replies in result. Each job is a full agent run that takes from half a minute to several minutes, so use it for work that needs judgement or writing. If the run is still going when the wait ends, the reply has wait_timed_out: true — then call wait_for_run with its id and never call ask_crevio again, which would do the job twice. Run ids do not expire, and only the user who started a run can read it.

Input Schema

TableJSON Schema

Name	Required	Description
`speed`	No	faster thinks less, for simple jobs where latency matters more than depth; smarter (default) thinks it through.
`message`	Yes	What you want done, in natural language, with all the context needed — the agent does not see your conversation.
`approval_mode`	No	autonomous (default) acts without review; supervised pauses in needs_input for your approval before each gated write (pending_approval_ids) and otherwise finishes with its proposal; read_only forbids writes.
`idempotency_key`	No	Always pass one, unique to this request. Retrying with the same key within an hour returns the run the first call started instead of doing the job twice.
`timeout_seconds`	No	How long to wait for the run to finish before returning. Defaults to, and is capped at, 300 s when your client accepts a streamed reply (text/event-stream), 90 s otherwise. The run keeps going when the wait ends first.

Output Schema

TableJSON Schema

Name	Required	Description
`id`	Yes
`object`	Yes
`result`	No
`reused`	No
`status`	Yes
`chat_id`	No
`summary`	No
`task_id`	Yes
`next_step`	No
`task_name`	No
`created_at`	Yes
`started_at`	No
`completed_at`	No
`error_message`	No
`wait_timed_out`	No	The wait ended first. The work is still running: follow it with the tool next_step names.
`credits_consumed`	No
`pending_approval_ids`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed7 schema fields changedSep 29, 2026

changedInput schema / properties / idempotency_key / description

Previous value: -"Retrying with the same key returns the run the first call started instead of starting another."New value: +"Always pass one, unique to this request. Retrying with the same key within an hour returns the run the first call started instead of doing the job twice."

changedInput schema / properties / message / description

Previous value: -"What you want done, in natural language."New value: +"What you want done, in natural language, with all the context needed — the agent does not see your conversation."

addedInput schema / properties / speed

Added value: +{
+  "description": "faster thinks less, for simple jobs where latency matters more than depth; smarter (default) thinks it through.",
+  "enum": [
+    "faster",
+    "smarter"
+  ],
+  "type": "string"
+}

changedInput schema / properties / timeout_seconds / description

Previous value: -"How long to wait for the run to settle before returning (default 60, max 90). On wait_timed_out, call wait_for_run with the run id."New value: +"How long to wait for the run to finish before returning. Defaults to, and is capped at, 300 s when your client accepts a streamed reply (text/event-stream), 90 s otherwise. The run keeps going when the wait ends first."

changedInput schema / properties / timeout_seconds / maximum
```
Previous value: -90New value: +300
```
addedOutput schema / properties / next_step
```
Added value: +{
+  "type": "string"
+}
```

addedOutput schema / properties / wait_timed_out

Added value: +{
+  "description": "The wait ended first. The work is still running: follow it with the tool next_step names.",
+  "type": "boolean"
+}

Changed1 schema field changedSep 24, 2026

changedInput schema / properties / approval_mode / description

Previous value: -"autonomous (default) acts without review; supervised pauses in needs_input for your review before finishing; read_only forbids writes."New value: +"autonomous (default) acts without review; supervised pauses in needs_input for your approval before each gated write (pending_approval_ids) and otherwise finishes with its proposal; read_only forbids writes."

Changed1 schema field changedSep 22, 2026

changedOutput schema / properties / status / enum

Previous value: -[
-  "pending",
-  "running",
-  "completed",
-  "failed",
-  "needs_input",
-  "timed_out",
-  "unknown"
-]New value: +[
+  "pending",
+  "running",
+  "completed",
+  "failed",
+  "needs_input",
+  "timed_out",
+  "unknown",
+  "skipped"
+]

Changed1 schema field changedSep 6, 2026

changedOutput schema / properties / status / enum

Previous value: -[
-  "pending",
-  "running",
-  "completed",
-  "failed",
-  "needs_input",
-  "timed_out"
-]New value: +[
+  "pending",
+  "running",
+  "completed",
+  "failed",
+  "needs_input",
+  "timed_out",
+  "unknown"
+]

Changed1 schema field changedSep 2, 2026

changedOutput schema / properties / status / enum

Previous value: -[
-  "pending",
-  "running",
-  "completed",
-  "failed",
-  "needs_input"
-]New value: +[
+  "pending",
+  "running",
+  "completed",
+  "failed",
+  "needs_input",
+  "timed_out"
+]

First observedAug 22, 2026

TDQS

A4.4/5.0

Behavior4/5

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

Adds substantial context beyond annotations: latency (half a minute to several minutes), the wait_timed_out signal, run-id permanence, and read-restriction to the run's initiator. The destructiveHint annotation is not directly explained, but the description's mention of approval modes elsewhere plus the timeout/retry semantics make the behavior clear enough. A point off only because it doesn't spell out the destructive/write implications that the annotation flags.

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?

Front-loads the core action and then layers the essential caveats (timeout, retry-once, run-id ownership) in a compact paragraph. Slightly dense in the second half but every clause carries operational value; nothing is decorative.

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?

Given a 5-param tool with an output schema and rich annotations, the description closes the remaining gaps: the wait/timeout flow, the double-execution hazard, and the ownership/expiry rules for run ids. An agent has everything needed to call it correctly and handle the timeout path.

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

Parameters3/5

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

Schema coverage is 100% and the schema descriptions are thorough (speed, approval_mode, idempotency_key, timeout_seconds). The description adds no parameter-level detail beyond reinforcing that 'result' carries the reply. Baseline 3 is appropriate when the schema already documents every parameter.

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+resource: 'Delegate a job to the Crevio agent and wait for the result.' It distinguishes itself from siblings like start_chat/get_run by framing itself as a full agent run that replies in `result`, and the follow-up routing to wait_for_run makes the boundary explicit.

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 says when to use it ('work that needs judgement or writing') and gives a hard when-not / what-to-do-instead rule: on wait_timed_out, 'call wait_for_run with its id and never call ask_crevio again, which would do the job twice.' That is exactly the alternative-routing guidance most definitions lack.

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

All Crevio tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?