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

Annotations already cover the safety profile (readOnly, idempotent, non-destructive, closed-world), so the description's job is to add interpretation context, and it does so substantively: it warns of selection bias (samples count improvements, not all attempts) and non-causality of before/after deltas. That is exactly the kind of caveat an agent needs to avoid overclaiming. It stops short of describing return shape or volume.

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?

Two tight sentences with no filler, and the caveats are packed efficiently with a semicolon. Slightly dense jargon in the first sentence keeps it from a 5, but nothing is wasted.

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 read-only tool with fully documented params and no output schema, the description should at least convey the shape of the evidence returned (per-channel breakdown? recommendations, as the limit parameter implies?). It covers the epistemic caveats well but leaves the return content underspecified.

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 limit and windowDays are already documented, giving a baseline of 3. However, the description's '14-day improvements' framing creates friction with the schema's windowDays options of 30 or 90, leaving the agent unsure whether the window is fixed at 14 days or the enum controls it. That ambiguity is a net negative for parameter understanding.

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 verb (Get) and resource (project channel evidence from closed 14-day improvements) are stated, but 'canonical closed 14-day improvements' is internal jargon that an agent cannot decode without prior knowledge of the product. It also lacks any differentiation from the many sibling read tools (get_direct_lift, get_outcomes_ledger, get_adoption_metrics) that could plausibly return similar evidence.

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?

There is no statement of when to call this versus alternatives, and no prerequisites or exclusions. The only guidance is interpretive ('samples count improvements, not all attempts'), which tells the agent how to read the data rather than when to fetch it.

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