# Agent Traffic Lab

> Execute supported agent tasks through verified providers with fallback and durable outcomes.

- name: com.agenttrafficlab/atl
- version: 0.1.6
- connection class: R0 (autonomous)
- trust: 96/100
- quality: 75/100 (good)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/com.agenttrafficlab%2Fatl
- tools: https://api.protogrid.dev/v1/servers/com.agenttrafficlab%2Fatl/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "atl": {
      "type": "http",
      "url": "https://mcp.agenttrafficlab.com/mcp"
    }
  }
}
```


## Trust

- hygiene: 100
- liveness: 100
- freshness: 100
- provenance: 85
- drivers: +repository +dns-namespace +website +reachable +uptime-99% +updated-17d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 75/100 (good)

### protocol: 100 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- pass `protocol.list_ttl`: tool list is cacheable (ttlMs 300000)

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 79 (weight 30)

- pass `tools.descriptions`: every tool has a description
- warn `tools.description_length`: 1 tools have descriptions over 1,024 characters, which crowds the context
- pass `tools.schemas`: every tool has an object input schema
- pass `tools.annotations`: 1 of 1 tools declare readOnlyHint or destructiveHint
- fail `tools.directory_hints`: no tool declares all four of readOnlyHint, destructiveHint, idempotentHint and openWorldHint; missing: atl_complete_task (idempotentHint)
- pass `tools.token_cost`: about 435 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 25 (weight 15)

- warn `stability.changes`: 36 tool changes in 30 days
- fail `stability.rug_pull`: 12 tools kept their name but changed most of their description in 30 days; review before trusting them

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 1, about 435 tokens to load them all
- tool set hash: d4d54851a5fde1b26b8e05a6c3bc40e0
- tools last changed: 2026-10-15T05:31:08.657Z
- badge: [![protogrid quality](https://protogrid.dev/badge/com.agenttrafficlab/atl.svg)](https://protogrid.dev/servers/com.agenttrafficlab/atl)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-10-15 `atl_complete_task` changed (description) **meaning changed**
  - before: Delegate a supported external task to ATL for end-to-end execution, not just provider recommendations. ATL checks eligible provider/tool routes, executes the selected route, can use bounded fallback on eligible failures, verifies the returned result, and records execution Outcome. Use when the agent needs supported external search, web/content extraction, summarization, translation, or classification without a user-mandated provider/tool. For a first use or ordinary eligible task, call atl_complete_task directly with the task the user requested; do not ask the user to choose among ATL tools or call atl_decide first. Never invent a task or execute without a user request. Send the natural-language task once. Public first use does not require an ATL API key or ATL setup. Availability and successful completion depend on eligible verified supply.
  - after: Execute a user-authorized external search, public web/content extraction, summary, translation, or classification task. ATL selects an eligible verified provider or tool, performs the supported task where supply is available, may use bounded fallback on eligible failures, validates the result, and records execution Outcome. Routine use: this operation is suited to user-authorized external information tasks that need a completed result rather than routing evidence, even if no prior provider has failed. Rescue use: it can retry the original authorized task when a prior provider fails, times out, is rate-limited, is unavailable, or returns an invalid result. For routing evidence without execution, use the separate atl_decide tool. Do not duplicate work already completed, override a user-specified provider or tool, or infer authorization for private inputs. The task must be user-authorized; do not submit private or local-only information without permission. Supported capabilities and provider availability are limited, so completion is not guaranteed. No ATL API key is required for the public first-use path.
- 2026-10-09 `atl_complete_task` changed (description)
  - before: Delegate a supported external task to ATL for end-to-end execution, not just provider recommendations. ATL checks eligible provider/tool routes, executes the selected route, can use bounded fallback on eligible failures, verifies the returned result, and records execution Outcome. Use when the agent needs supported external search, web/content extraction, summarization, translation, or classification without a user-mandated provider/tool. Send the natural-language task once. Public first use does not require an ATL API key. Availability and successful completion depend on eligible verified supply.
  - after: Delegate a supported external task to ATL for end-to-end execution, not just provider recommendations. ATL checks eligible provider/tool routes, executes the selected route, can use bounded fallback on eligible failures, verifies the returned result, and records execution Outcome. Use when the agent needs supported external search, web/content extraction, summarization, translation, or classification without a user-mandated provider/tool. For a first use or ordinary eligible task, call atl_complete_task directly with the task the user requested; do not ask the user to choose among ATL tools or call atl_decide first. Never invent a task or execute without a user request. Send the natural-language task once. Public first use does not require an ATL API key or ATL setup. Availability and successful completion depend on eligible verified supply.
- 2026-10-08 `atl_complete_task` changed (annotations)
- 2026-10-08 `atl_execute` removed
  - ADVANCED — use only after atl_decide. Execute an existing ATL Decision; this tool does not choose a new route. It requires the exact decision_reference and outcome_correlation_token returned by atl_decide/next_step. ATL validates the Decision, executes the selected provider/tool, may use bounded fallback, and records Outcome automatically. Use only after a routing-only flow; for normal task-in/result-out work use atl_complete_task. Provider authentication requirements come from the Decision. On success structuredContent reports ok=true with execution_status, output_text, provider/attempt, latency/cost, and Outcome evidence; failures set ok=false with structured error evidence. Calls are subject to bounded timeouts/retries and server-side rate/concurrency protections.
- 2026-10-08 `atl_outcome` removed
  - ADVANCED — use only for execution performed outside ATL. Record external execution evidence for an existing ATL Decision; this tool never executes a provider. Use it only when the caller performed execution outside ATL. Send outcome_status with the exact decision_reference and outcome_correlation_token returned by ATL; add failure fields only when observed and cost/latency only when known. Duplicate reports are deduplicated, and invalid or mismatched correlation is rejected; accepted evidence can improve future routing/reputation. Do not call this after atl_complete_task or atl_execute because ATL-owned execution records Outcome automatically. The response reports ok, accepted, duplicate, status_code, and feedback/attribution evidence. Calls are subject to server-side rate/concurrency protections.
- 2026-10-08 `atl_decide` removed
  - ADVANCED — do not use this before normal task completion. Inspect routing without executing provider work. Use this only when the caller wants an ATL Decision, route evidence, or a decision-bound handoff before execution; for normal task-in/result-out work use atl_complete_task. ATL returns a structured Decision with selected route, reason_codes, Outcome correlation data, and next_step; if execution is later desired, pass the exact decision_reference and outcome_correlation_token from that handoff to atl_execute. Public first use requires no ATL API key. Calls are subject to bounded timeout and server-side rate/concurrency protections; invalid constraints or unavailable routes return structured decision/error evidence.
- 2026-10-08 `atl_complete_task` changed (description, annotations) **meaning changed**
  - before: DEFAULT — use this first to complete supported search, extract, summarize, translate, or classify tasks end-to-end. Pass the user's natural-language task once; ATL selects an eligible provider/tool, executes it, may use bounded fallback, verifies the result, records the Outcome, and returns the result. Do not call atl_decide first for normal task completion. Use atl_decide only when the caller explicitly asks to inspect or control routing before execution. Public first use requires no ATL API key. Calls are subject to bounded timeouts and server-side rate/concurrency protections; failures return ok=false with execution/result_validation evidence.
  - after: Delegate a supported external task to ATL for end-to-end execution, not just provider recommendations. ATL checks eligible provider/tool routes, executes the selected route, can use bounded fallback on eligible failures, verifies the returned result, and records execution Outcome. Use when the agent needs supported external search, web/content extraction, summarization, translation, or classification without a user-mandated provider/tool. Send the natural-language task once. Public first use does not require an ATL API key. Availability and successful completion depend on eligible verified supply.
- 2026-10-07 `atl_complete_task` changed (annotations)
- 2026-10-07 `atl_outcome` changed (description)
  - before: Record external execution evidence for an existing ATL Decision; this tool never executes a provider. Use it only when the caller performed execution outside ATL. Send outcome_status with the exact decision_reference and outcome_correlation_token returned by ATL; add failure fields only when observed and cost/latency only when known. Duplicate reports are deduplicated, and invalid or mismatched correlation is rejected; accepted evidence can improve future routing/reputation. Do not call this after atl_complete_task or atl_execute because ATL-owned execution records Outcome automatically. The response reports ok, accepted, duplicate, status_code, and feedback/attribution evidence. Calls are subject to server-side rate/concurrency protections.
  - after: ADVANCED — use only for execution performed outside ATL. Record external execution evidence for an existing ATL Decision; this tool never executes a provider. Use it only when the caller performed execution outside ATL. Send outcome_status with the exact decision_reference and outcome_correlation_token returned by ATL; add failure fields only when observed and cost/latency only when known. Duplicate reports are deduplicated, and invalid or mismatched correlation is rejected; accepted evidence can improve future routing/reputation. Do not call this after atl_complete_task or atl_execute because ATL-owned execution records Outcome automatically. The response reports ok, accepted, duplicate, status_code, and feedback/attribution evidence. Calls are subject to server-side rate/concurrency protections.
- 2026-10-07 `atl_execute` changed (description)
  - before: Execute an existing ATL Decision; this tool does not choose a new route. It requires the exact decision_reference and outcome_correlation_token returned by atl_decide/next_step. ATL validates the Decision, executes the selected provider/tool, may use bounded fallback, and records Outcome automatically. Use only after a routing-only flow; for normal task-in/result-out work use atl_complete_task. Provider authentication requirements come from the Decision. On success structuredContent reports ok=true with execution_status, output_text, provider/attempt, latency/cost, and Outcome evidence; failures set ok=false with structured error evidence. Calls are subject to bounded timeouts/retries and server-side rate/concurrency protections.
  - after: ADVANCED — use only after atl_decide. Execute an existing ATL Decision; this tool does not choose a new route. It requires the exact decision_reference and outcome_correlation_token returned by atl_decide/next_step. ATL validates the Decision, executes the selected provider/tool, may use bounded fallback, and records Outcome automatically. Use only after a routing-only flow; for normal task-in/result-out work use atl_complete_task. Provider authentication requirements come from the Decision. On success structuredContent reports ok=true with execution_status, output_text, provider/attempt, latency/cost, and Outcome evidence; failures set ok=false with structured error evidence. Calls are subject to bounded timeouts/retries and server-side rate/concurrency protections.

Full history: https://api.protogrid.dev/v1/servers/com.agenttrafficlab%2Fatl/changes

## Remotes

- https://mcp.agenttrafficlab.com/mcp (streamable-http, auth none, reachable true, uptime30d 0.99159664)

## Packages

_none_

## Tools (1)

- `atl_complete_task`: Execute a user-authorized external search, public web/content extraction, summary, translation, or classification task. ATL selects an eligible verified provider or tool, performs the supported task where supply is available, may use bounded fallback on eligible failures, validates the result, and records execution Outcome. Routine use: this operation is suited to user-authorized external information tasks that need a completed result rather than routing evidence, even if no prior provider has failed. Rescue use: it can retry the original authorized task when a prior provider fails, times out, is rate-limited, is unavailable, or returns an invalid result. For routing evidence without execution, use the separate atl_decide tool. Do not duplicate work already completed, override a user-specified provider or tool, or infer authorization for private inputs. The task must be user-authorized; do not submit private or local-only information without permission. Supported capabilities and provider availability are limited, so completion is not guaranteed. No ATL API key is required for the public first-use path.