swarme_experiment_createRun 2–6 compatible models against one canonical brief with a credit budget.
| Name | Required | Description | Default |
|---|---|---|---|
| task | Yes | ||
| blind | No | ||
| input | Yes | ||
| models | Yes | ||
| objective | No | ||
| budget_units | No |
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must carry the full behavioral burden, but it only states that models are run against a brief with a credit budget. It does not disclose whether execution is asynchronous, whether credit is consumed immediately or only on completion, what 'compatible' means, or what side effects occur.
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?
The description is a single, front-loaded sentence with no filler words; it earns its place. It is concise to the point of under-specification, but that is better penalized under other dimensions rather than structure.
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?
For a 6-parameter tool with nested objects, two enums, and no annotations, the description is far from complete. It fails to clarify model compatibility, objective semantics, blind mode, or budget behavior; although an output schema exists, the input side remains inadequately described.
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?
Schema description coverage is 0%, so the description must compensate, but it only loosely maps to 'models' (2–6 models) and 'budget_units' (credit budget). It does not explain the 'task' enum, the 'input' object structure, 'blind', or 'objective', leaving most of the 6 parameters semantically underdefined.
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 uses a specific verb and resource: 'Run 2–6 compatible models against one canonical brief with a credit budget.' This clearly conveys the tool's multi-model experiment purpose and differentiates it from single-run sibling tools like swarme_ai_run or swarme_tool_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 implies the tool is for running multi-model experiments with a budget, but it does not explicitly say when to prefer this tool over alternatives or mention exclusions such as single-model runs or non-experiment workflows. The usage context is present but left largely to inference.
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.
Tools are grouped into distinct resource families (account, AI, capabilities, discover, experiments, tool runs, uploads, vault), so most are easy to tell apart. However, multiple 'status' tools and the two Vault metadata tools (documents and summary) could be confused without reading the descriptions carefully.
All tools share a consistent swarme_ prefix and snake_case style, making them predictable. The pattern is not uniformly verb_noun, and there is a plural mismatch between swarme_capabilities_search and swarme_capability_describe, but the naming is generally coherent.
22 tools is on the heavy side and above the typical 3-15 range for a focused MCP server. While the platform covers many domains and each tool has a role, several status-related tools could have been consolidated.
Core workflows like tool runs, experiments, AI model jobs, and Vault fills are well covered. Notable gaps remain: no cancellation for AI jobs or experiments, no listing endpoints for runs/experiments, no Vault field read/delete, and the upload session has no companion completion/status tool.