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

Annotations carry only generic flags (readOnlyHint=false, idempotentHint=false, openWorldHint=true), so the description carries the behavioral burden and delivers. It discloses that runs take minutes to hours and a 'running' response is normal rather than an error, that NEEDS_REVIEW means paused for human review in the dashboard, and that the workflow must be deployed before it is runnable. It also adds the outputs/package exclusion and directs the agent to follow llmContext guidance in results.

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 dense sentences, each carrying operational weight: purpose, deployment prerequisite, async/status semantics, the package alternative, outputs injection, and llmContext guidance. Purpose is front-loaded and there is no filler, though the trailing llmContext clause is a slightly tacked-on afterthought and the density makes it a demanding read for an 11-parameter tool.

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 high-complexity tool with nested objects, mutual exclusions, and an output schema, the description covers the operational essentials: deployment gating, async expectations, special status meanings, mode selection, and the draft version escape hatch. The output schema handles return-value documentation and the schema handles parameter detail, so nothing an agent needs to invoke correctly is missing.

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 100%, and every parameter is already richly documented, including mutual-exclusivity constraints on file/package/outputs and environment restrictions tied to get_me. The description adds some semantic rationale (why outputs exists: to skip recomputation; package for merged multi-file runs) but mostly restates what the schema already covers, so the marginal value over the schema is modest.

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+resource: 'Start a run of a deployed multi-step workflow on a document.' It anchors the tool in the 'workflows group' and its single-run scope distinguishes it from siblings like run_workflow_batch, get_workflow_run, and cancel_workflow_run. The file-vs-package modes further delimit the tool's operation.

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?

Provides concrete when-to-use context: a workflow is not runnable until a version is deployed (or pass 'draft'), and for one merged run over 2-50 related files you should use package instead of file. The async status note implies the agent should poll elsewhere, but it never explicitly names alternative tools (e.g., run_workflow_batch or get_workflow_run), so routing guidance is clear but not fully explicit.

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