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

Annotations already declare destructiveHint=true, idempotentHint=true and openWorldHint=true, so safety framing is partly covered; the description nonetheless adds real context: human approval is a hard prerequisite, current CRM delivery and source lead grants must still be in place at call time, and operationId should be reused after an uncertain outcome, which operationalizes the idempotency hint. The side-effect boundary (no messages, no subscription changes) is useful scope disclosure. It stops short of saying what a successful delivery produces, but an output schema exists.

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 short sentences, front-loaded with the gating condition and with zero filler. The phrase 'Current CRM delivery and source lead grants remain required' is slightly telegraphic, compressing a permission check into an oblique clause, but no sentence is redundant.

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 destructive, open-world, idempotent mutation with a nested payload and six parameters, the description covers preconditions, retry behavior, and scope limits, and return values are handled by the existing output schema. The remaining shortfall is the unexplained envelope parameters, which leaves an agent guessing about work_id and request_key on a call that performs a real CRM delivery.

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 only 50% and the nested request object's fields carry no descriptions, so the description does meaningful work: it fixes action='deliver', marks connectionId as the reviewed connection, ties reviewId and fingerprint to the human review, and gives operationId an uncertainty-retry semantic. The unaddressed gap is the top-level envelope, notably work_id and request_key, which are never explained anywhere.

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: submitting a human-approved CRM lead delivery, with the concrete action enum value 'deliver' named. The closing clause 'this does not send messages or manage subscriptions' fences the tool off from adjacent GenAds messaging/subscription tools, though it never names a specific sibling such as genads_crm_review or genads_crm_request.

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 opening restriction ('only an exact human-approved CRM delivery') tells the agent the precondition for calling at all, and 'Reuse operationId after uncertainty' prescribes the retry pattern. It also states the exclusion of messaging and subscription management. What is missing is an explicit pointer to where reviewId/fingerprint must be obtained (the review tools), so the agent must infer the prerequisite chain.

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