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

Annotations already declare readOnly/idempotent/non-destructive, and the description still adds substantial context the annotations cannot: the statistical basis (mSPRT adjusted for arm count, valid under repeated peeking), the exact verdict vocabulary, the 'not enabled' degenerate response, the credential dependency, and the 403 refusal path for self-registered sandbox keys.

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 purpose is front-loaded in the first clause, but the body is two long run-on sentences that pack the statistical method, bandit floors, weight history and two auth/claim failure modes into undifferentiated prose. Much of it is genuinely useful, yet a bulleted or segmented layout would make the return payload scannable rather than a wall of comma-spliced clauses.

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?

No output schema exists, so the description carries the full burden of describing return values — and it does, covering verdict, reason, headline, per-arm win probability, expected loss, traffic split, bandit weights/floors and history. Combined with the auth and 'not enabled' states, an agent has everything needed to call it and interpret the result.

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 both parameters are already documented, making 3 the baseline. The description earns a notch above by clarifying what history_limit actually counts ('every change the bandit made and every change to its settings'), which tells the agent the history feed is a union of two record types rather than a single log.

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?

Opens with a specific verb and resource: 'Read whether one of your stored experiments can stop', and enumerates the verdict, per-arm chance to beat control, traffic split, bandit state and weight history. The scope is unmistakable against generic readers like get_experiment_definition, though it never names the nearest sibling (check_variant_significance or get_experiment_results) to draw the line explicitly.

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 concrete preconditions that gate invocation: sequential testing must be switched on (otherwise 'not enabled'), a Last Price credential for the owning workspace is required, and an unclaimed sandbox workspace is refused with 403. It stops short of stating when to prefer this over get_experiment_results or check_variant_significance, so the when-not-to-use side is inferred.

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