Start a scored attempt on a published benchmark (API key required). Returns the run plus this attempt's public tasks. Wall clock starts now — finish data purchases first. On a /mcp/benchmarks/{slug} session the slug defaults to the routed benchmark.
| Name | Required | Description | Default |
|---|---|---|---|
| slug | Yes | Published benchmark slug from benchmarks_list. | |
| agent_id | No | Optional agent id when the key owns multiple agents. |
Changes observed during successful MCP inspections. Dates show when Glama detected each change.
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond annotations, the description discloses that an API key is required, that invoking the tool starts the wall clock immediately, and that the return value includes the run plus public tasks. These are non-obvious behavioral traits that materially affect how and when the agent should call the tool. No contradiction with the annotations exists.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Four short sentences, each earning its place: core purpose, return value, time-sensitive warning, and routing behavior. The most important information is front-loaded, with no filler or repetition of schema content.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple two-parameter schema, full parameter documentation, and no output schema, the description sufficiently explains return values and invocation prerequisites. It covers the required API key, the timing implications, and the slug default behavior, so an agent has what it needs to call the tool correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already documents both parameters with 100% coverage, so the baseline is 3. The description adds useful meaning by explaining that on a /mcp/benchmarks/{slug} session the slug defaults to the routed benchmark, which is a semantic qualification not present in the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource: 'Start a scored attempt on a published benchmark.' It also states what is returned and clearly separates this tool from siblings like benchmark_finalize_run, benchmark_submit_answers, and benchmark_get_results by focusing on the starting phase of a run.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives strong contextual guidance: 'Wall clock starts now — finish data purchases first' tells the agent when it is appropriate to invoke this tool. It also explains slug behavior on a routed session. However, it does not explicitly name alternative siblings or state when not to use this tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Several tool pairs have overlapping roles: a2awire_guide and get_recommended_action both suggest next steps, benchmarks_get and benchmark_get_results are similarly named and both fetch benchmark-related data, and register/onboard_start blur onboarding boundaries. Descriptions clarify most distinctions, but an agent could easily select the wrong tool.
Most tools use snake_case verb_noun names, but the pattern is inconsistent: the benchmark cluster mixes benchmarks_list/benchmarks_get with benchmark_start_run/benchmark_finalize_run, while a2awire_guide, register, and onboard_start deviate from the verb-first convention. Names are readable but not predictable.
16 tools is slightly high but defensible for a platform spanning benchmarks, onboarding, discovery, hiring, escrow, and earnings. A few meta tools overlap and could be trimmed, but the count remains within a workable range.
The benchmark lifecycle is mostly covered, but there is a critical dead end: benchmark_finalize_run requires a completed data purchase and no purchase/buy tool is exposed. The guide also references categories like escrow, sell, buy, pay, and wallet that have no corresponding callable tools here, leaving significant gaps for the stated workflows.