Agent Traffic Lab

atl_decide

Find the right tool, MCP server, or provider to execute a task. Call this when you need a tool for a task, need to find a provider, need to choose an MCP server, are unsure which tool or provider to use, need a reliable executable provider, have multiple possible providers or tools, hit a tool or provider failure, or need an alternative, fallback, or a way to route a task to the right capability. Just pass the task in plain natural language; no other fields are needed. After you execute the selected provider/tool, report the real result with atl_outcome.

Input Schema

TableJSON Schema

Name	Required	Description
`task`	Yes	Natural-language task. ATL infers the required capability when possible.
`timeout_ms`	No	Optional time limit for the decision request in milliseconds.
`constraints`	No	Optional hard requirements the selected provider must meet.
`max_retries`	No	Optional maximum retries allowed for the decision request.
`preferences`	No	Optional trade-offs such as cost, latency, or reliability.
`auth_available`	No	Whether the caller can authenticate to the selected provider.
`desired_outcome`	No	Optional result the caller wants from the task.
`required_capability`	No	Capability the selected provider must support.

Schema Changelog

Changes observed during successful MCP inspections.

Changed7 schema fields changedAug 17, 2026

removedInput schema / properties / caller_id

Removed value: -{
-  "description": "Optional caller identity for the request.",
-  "type": "string"
-}

removedInput schema / properties / correlation_id

Removed value: -{
-  "description": "Caller correlation ID shared across related work.",
-  "type": "string"
-}

removedInput schema / properties / idempotency_key

Removed value: -{
-  "description": "Optional key for safely deduplicating repeated decision requests.",
-  "type": "string"
-}

addedInput schema / properties / max_retries / maximum
```
Added value: +2
```

removedInput schema / properties / request_id

Removed value: -{
-  "description": "Caller request ID for tracing this decision.",
-  "type": "string"
-}

removedInput schema / properties / session_id

Removed value: -{
-  "description": "MCP session ID returned by initialize.",
-  "type": "string"
-}

removedInput schema / properties / tenant_id

Removed value: -{
-  "description": "Optional tenant identity for the request.",
-  "type": "string"
-}

Changed13 schema fields changedAug 13, 2026

addedInput schema / properties / auth_available / description

Added value: +"Whether the caller can authenticate to the selected provider."

addedInput schema / properties / caller_id / description

Added value: +"Optional caller identity for the request."

addedInput schema / properties / constraints / description

Added value: +"Optional hard requirements the selected provider must meet."

addedInput schema / properties / correlation_id / description

Added value: +"Caller correlation ID shared across related work."

addedInput schema / properties / desired_outcome / description

Added value: +"Optional result the caller wants from the task."

addedInput schema / properties / idempotency_key / description

Added value: +"Optional key for safely deduplicating repeated decision requests."

addedInput schema / properties / max_retries / description

Added value: +"Optional maximum retries allowed for the decision request."

addedInput schema / properties / preferences / description

Added value: +"Optional trade-offs such as cost, latency, or reliability."

addedInput schema / properties / request_id / description

Added value: +"Caller request ID for tracing this decision."

addedInput schema / properties / required_capability / description
```
Added value: +"Capability the selected provider must support."
```

addedInput schema / properties / session_id / description

Added value: +"MCP session ID returned by initialize."

addedInput schema / properties / tenant_id / description

Added value: +"Optional tenant identity for the request."

addedInput schema / properties / timeout_ms / description

Added value: +"Optional time limit for the decision request in milliseconds."

Changed1 schema field changedAug 13, 2026

addedInput schema / properties / task / description

Added value: +"Natural-language task. ATL infers the required capability when possible."

First observedAug 12, 2026

TDQS

A4.3/5.0

Behavior4/5

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

No annotations are provided, so the description carries the burden. It explains that only the plain-language task is required ('no other fields are needed'), that it selects a provider/tool for the agent to execute rather than executing itself, and that results should be reported via atl_outcome. This gives a clear behavioral picture, though it doesn't describe return format or failure modes.

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?

Four sentences, front-loaded with the main purpose. The long list of trigger conditions is exhaustive but still readable and earns its place; not overly wordy.

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?

For an 8-parameter tool with no output schema and no annotations, the description covers purpose, when to use, what input to pass, and the after-action (atl_outcome). It does not detail return values or optional parameter usage, but the schema covers the latter and the workflow is clear.

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% (all 8 parameters documented). Per the rubric, baseline is 3. The description adds the key note that 'no other fields are needed' and the task suffices, but does not add detail beyond the schema for the optional parameters. This matches the baseline.

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 description explicitly states the tool's purpose: 'Find the right tool, MCP server, or provider to execute a task.' It uses a specific verb ('Find') and clear resource types, and distinguishes from sibling atl_outcome by directing reporting after execution.

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?

The description provides an extensive list of 'Call this when' triggers (need a tool, provider, MCP server, unsure, multiple options, failure, fallback/routing). It also clarifies what to do after: 'After you execute the selected provider/tool, report the real result with atl_outcome,' which guides alternative usage.

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

All Agent Traffic Lab 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?