Chat Messages Validator

Overview

You send 1 to 1,000 messages with roles SYSTEM, DEVELOPER, USER, ASSISTANT or TOOL. We check empty content, duplicate tool-call ids, orphan TOOL results, unresolved assistant calls and RFC 8259 argumentsJson, under profile GENERIC_V1 only. tokensEstimated is always false — we do not count tokens, moderate content, or call a model. Multimodal parts are recorded as metadata and never fetched.

A tool result with no matching assistant call, or two calls sharing an id, looks like a normal transcript until the next model call fails in a way you cannot replay.

When to use it

Scope

Non-goals

• Does not call a model or apply provider token limits
• Does not moderate or rewrite the transcript
• Does not fetch multimodal assets named in content parts

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"ai.chat_messages_validate","input":{"messages":[{"role":"USER","content":"Hello"}]}}'

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("ai.chat_messages_validate", {"messages":[{"role":"SYSTEM","content":"You are a support agent."},{"role":"USER","content":"Where is order NW-1?"},{"toolCalls":[{"name":"lookup_order","argumentsJson":"{\"orderId\":\"NW-1\"}","id":"call_1"}],"content":"Let me look that up.","role":"ASSISTANT"},{"toolCallId":"call_1","content":"{\"status\":\"shipped\"}","role":"TOOL"}]});

Errors

Other limits

Limitations

• Does not call a model or apply provider token limits
• Does not moderate or rewrite the transcript
• Does not fetch multimodal assets named in content parts
• Counting tokens or pricing a completion
• Safety moderation

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "ai.chat_messages_validate",
  "input": {
    "messages": [
      {
        "role": "SYSTEM",
        "content": "You are a support agent."
      },
      {
        "role": "USER",
        "content": "Where is order NW-1?"
      },
      {
        "toolCalls": [
          {
            "name": "lookup_order",
            "argumentsJson": "{\"orderId\":\"NW-1\"}",
            "id": "call_1"
          }
        ],
        "content": "Let me look that up.",
        "role": "ASSISTANT"
      },
      {
        "toolCallId": "call_1",
        "content": "{\"status\":\"shipped\"}",
        "role": "TOOL"
      }
    ]
  }
}

Examples

{
  "messages": [
    {
      "role": "SYSTEM",
      "content": "You are a support agent."
    },
    {
      "role": "USER",
      "content": "Where is order NW-1?"
    },
    {
      "toolCalls": [
        {
          "name": "lookup_order",
          "argumentsJson": "{\"orderId\":\"NW-1\"}",
          "id": "call_1"
        }
      ],
      "content": "Let me look that up.",
      "role": "ASSISTANT"
    },
    {
      "toolCallId": "call_1",
      "content": "{\"status\":\"shipped\"}",
      "role": "TOOL"
    }
  ]
}

{
  "profile": "GENERIC_V1",
  "valid": true,
  "toolCallCount": 1,
  "messageCount": 4,
  "diagnostics": [],
  "tokensEstimated": false,
  "roleCounts": {
    "ASSISTANT": 1,
    "TOOL": 1,
    "SYSTEM": 1,
    "USER": 1,
    "DEVELOPER": 0
  }
}

{
  "messages": [
    {
      "role": "USER",
      "content": "Hello"
    },
    {
      "toolCallId": "missing",
      "content": "{}",
      "role": "TOOL"
    }
  ]
}

{
  "profile": "GENERIC_V1",
  "valid": false,
  "toolCallCount": 0,
  "messageCount": 2,
  "diagnostics": [
    {
      "code": "ORPHAN_TOOL_RESULT",
      "severity": "ERROR",
      "message": "A TOOL turn does not resolve any assistant tool call id.",
      "path": "/messages/1/toolCallId",
      "source": "squillify",
      "details": {}
    }
  ],
  "tokensEstimated": false,
  "roleCounts": {
    "ASSISTANT": 0,
    "TOOL": 1,
    "SYSTEM": 0,
    "USER": 1,
    "DEVELOPER": 0
  }
}

Recommended use cases

• Check an assistant tool call is followed by a matching TOOL turn
• Reject duplicate tool-call ids before you send the transcript on

Avoid for

• Counting tokens or pricing a completion
• Safety moderation

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags