CRM Connector

Propose: Send a message

propose_send_message

AI Agents (inside one system): Use when the person asks for this in a system whose AI agent may change records: An email or text message sent to a contact from the person's own system, as part of the conversation with them. The contact receives it, and only the person can approve it. Give one item per record, at most 25. Nothing is written yet: the answer shows, per record, what it holds now and would hold after, plus a proposal id. Show it to the person, say so first when something would be sent to the people in their system, ask whether to go ahead, and only then call apply_proposal; when needsApproval is true, share the approve link instead. Not for reading (use search_records or get_record).

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "system": "<system>",
      "items": "<items>"
    },
    "name": "propose_send_message"
  }
}

Input Schema

Name	Required	Description	Default
`items`	Yes	One object per record, at most 25.
`system`	Yes	Which AI agent: the system's id from get_ai_agent_options, or the name of the system or its connection. Ask the person which AI agent they mean unless they named it in this conversation; never pick one yourself.

Output Schema

Name	Required	Description	Default
`notice`	No
`changes`	Yes
`systemId`	Yes
`warnings`	Yes
`expiresAt`	Yes
`approveUrl`	Yes
`proposalId`	Yes
`systemName`	Yes
`needsApproval`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed4 schema fields changedNov 10, 2026

changedInput schema / properties / items / items / properties / channel / description

Previous value: -"How it is sent."New value: +"How it is sent; a reply on a ticket goes by email."

changedInput schema / properties / items / items / properties / id / description

Previous value: -"The id of the contact in the system."New value: +"The id of the contact or ticket in the system."

changedInput schema / properties / items / items / properties / type / description

Previous value: -"The type of the record: a contact."New value: +"The type of the record: a contact, or a ticket for a reply on the ticket itself."

changedInput schema / properties / items / items / properties / type / enum

Previous value: -[
-  "contact"
-]New value: +[
+  "contact",
+  "ticket"
+]

Changed1 schema field changedNov 9, 2026

addedInput schema / properties / items / items / properties / statement

Added value: +{
+  "description": "True to attach the contact's account statement to an email, where the system keeps one.",
+  "type": "boolean"
+}

Changed1 schema field changedNov 5, 2026

changedInput schema / properties / system / description

Previous value: -"Which AI agent: the system's id from get_ai_agents_info, or the name of the system or its connection. Ask the person which AI agent they mean unless they named it in this conversation; never pick one yourself."New value: +"Which AI agent: the system's id from get_ai_agent_options, or the name of the system or its connection. Ask the person which AI agent they mean unless they named it in this conversation; never pick one yourself."

AddedNov 5, 2026

TDQS

A4/5.0

Behavior3/5

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

Annotations already declare readOnlyHint=false, destructiveHint=false, openWorldHint=true and idempotentHint=false. The definition still adds substantial context beyond them: it is a no-write proposal that returns current vs. prospective state plus a proposal id, requires human approval, and caps input at 25 items. That extra disclosure is meaningful, though the score reflects that the safety profile is largely carried by annotations.

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

Conciseness3/5

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

Every sentence carries relevant information, but the opening clause (“AI Agents (inside one system): Use when the person asks for this in a system whose AI agent may change records”) is convoluted and delays the purpose statement to the second sentence. The approval workflow is useful but runs together with the read-alternative note in a single dense block.

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

Completeness4/5

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

With an output schema, full annotation coverage and 100% schema descriptions, the definition supplies what's missing on its side: the human-in-the-loop workflow, the proposal id, and the approve-link branch. The only modest gap is that nothing describes what happens on partial failure across the multi-item batch.

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 description coverage is 100%, so items fields, channel enums, send_at, subject and statement are already fully documented in the schema. The description adds little beyond restating the one-item-per-record and 25-item bound that maxItems already encodes, so the baseline of 3 applies.

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?

The definition states a specific verb+resource (“An email or text message sent to a contact from the person's own system”) and makes the dry-run nature explicit (“Nothing is written yet”). It also names the siblings it is not: search_records/get_record for reading and apply_proposal for committing, so an agent can place it without opening a schema.

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?

It gives an explicit trigger (use when the person asks for this in a system whose AI agent may change records), an explicit exclusion (“Not for reading”) with named alternatives, and a full escalation path: show the person the result, ask, then call apply_proposal, or share the approve link when needsApproval is true.

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

All CRM Connector 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?