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

Beyond annotations (readOnly=false, openWorld=true, destructive=false), the description discloses the key behavioral trait an agent cannot infer: the call is fire-and-forget, the run continues server-side, and results appear live on the task's page. It also prescribes post-call behavior ('Tell the user it has started and what it does next').

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?

Front-loads the asynchronous contract in the first clause, then layers input guidance and status-check guidance in order of usefulness. Slightly dense—the 'Tell the user it has started and what it does next' instruction is agent-facing coaching rather than tool semantics, though it does earn its place as orchestration guidance.

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 mutation tool with no output schema, it covers what happens after invocation, where to observe results, and how to supply input. The needs_confirmation/approvalId recovery path is left to the schema description rather than the tool description, which is a minor gap for an agent that may hit that response mid-flow.

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 coverage is 100%, so baseline is 3; the description still adds meaning by enumerating acceptable `input` forms ('pasted text, extracted attachment text, or JSON') and tying the parameter to a trigger having an input. It adds nothing extra for workspace or approvalId, but the schema already documents those.

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?

States a specific verb and resource ('Kicks off a run of a Task') and immediately defines its scope with 'returns at once; the run carries on by itself'. This asynchronous framing distinguishes it from siblings like workbench_tasks_run_view or workbench_flows_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?

Gives explicit conditions for passing `input` ('when the task expects an input document (its trigger has an input)'), names the alternative for status checks (workbench_tasks_run_view), and adds an explicit exclusion ('don't poll it in a loop'). When/when-not/alternative are all present.

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