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

Annotations already signal readOnlyHint=false (writes), openWorldHint=true (may trigger arbitrary downstream behavior), idempotentHint=false, destructiveHint=false. The description adds the key real-world behavioral fact: the tool returns a batchId immediately rather than waiting for completion, and progress must be polled via a different endpoint. It also flags that results may include llmContext guidance the agent should follow. These are not inferable from annotations, so this adds genuine value. Minor gap: doesn't mention failure semantics or partial-failure behavior, but the coverage is solid.

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?

Three sentences, zero fluff. The critical operational fact (no batch-get endpoint; use list_workflow_runs by batchId) is front-loaded in the second sentence, and the llmContext note is the only trailing detail. Every sentence earns its place, and the structure is ideal for an agent scanning fast.

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?

The description is complete for a batch-submission tool of moderate complexity. With an output schema present, the agent already knows the return shape. The description covers the key operational flow (submit → get batchId → poll list endpoint), and the schema covers parameters thoroughly. It doesn't explain retry, throttling, or partial failure, and it doesn't mention whether results arrive asynchronously per-run, but no output schema exists to confirm results structure. Given openWorldHint=true, a small note about side effects would have pushed this to 5.

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 per the rubric the baseline is 3. The description adds a small but valuable amount: it clarifies that the 'file' object requires exactly one of id/url/text and must not be a bare string, which is a subtle gotcha not fully obvious from the schema alone. It also summarizes the inputs array as '1-1000 documents.' The description does not fully restate the environment/workspace constraints, but the schema's get_me references already cover those. The description's added value nudges this above baseline.

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 states a specific verb+resource ('Submit up to 1,000 documents as one batch of workflow runs') and clearly distinguishes it from run_workflow (single run) and list_workflow_runs (progress tracking). It names the missing batch-get endpoint, which fully disambiguates the tool from get_workflow_run and list_workflow_runs.

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 explicitly states when to use this tool (for batching 1-1,000 documents) and, crucially, tells the agent what NOT to expect: 'workflow batches have NO batch-get endpoint — track progress with list_workflow_runs filtered by batchId.' This routes the agent to the correct sibling for follow-up, which is exactly the kind of when-to-use vs alternative guidance that's needed.

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