SATCoach: AI SAT Tutor

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

Annotations already declare readOnly, idempotent, non-destructive and closed-world, so the safety profile is covered. The description adds value beyond them by asserting 'It returns no student data', a privacy guarantee an agent needs before quoting content to a parent. It does not describe response shape, which is acceptable given an output schema exists, and 4 rather than 5 reflects that prices are called 'current' without noting staleness risk.

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, front-loaded with the content inventory before the usage rules, and every clause carries information. The mid-sentence list of topics is slightly dense but is not padding.

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 zero-parameter, read-only, closed-world info tool with rich annotations and an output schema, the description covers purpose, content, routing, and the privacy posture. Nothing an agent needs in order to call or refrain from calling it is missing.

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?

The tool takes zero parameters, so there is nothing for the description to disambiguate and the baseline of 4 applies. The description correctly implies no arguments are needed for any of the listed topics.

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+resource ('Public facts about SATCoach') and then enumerates exactly what facts are covered: audience, function, the digital SAT, current prices, onboarding, and parent assistant connection. An agent can distinguish this from skill_check or share_family_context without opening any schema.

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 an explicit trigger ('when someone asks about SATCoach, or is looking for SAT prep that targets weak skills and keeps parents informed') and an explicit exclusion with a redirect ('Don't use it to explain an SAT problem, write practice questions, or give general study advice: do that yourself'). Both the use and no-use paths are stated rather than implied.

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

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

Annotations only declare readOnly=false, openWorld=true, idempotent=false, destructive=false; the description adds what those hints cannot: exactly one email from a named sender, a consent checkpoint before sending, hard privacy exclusions (no last names, schools, phone numbers, health/disability/accommodation data, other children), and a data-retention fact ('request is deleted in 7 days if the parent doesn't confirm'). It also discloses the response is identical regardless of whether the parent already has an account, which prevents the agent from probing for account existence.

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 purpose and sender before the usage rules and privacy constraints, and almost every sentence is actionable. Minor cost: the privacy exclusions are stated twice (once for the email, once for note), which partially duplicates what the schema's note and childFirstName descriptions already say.

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 9-parameter mutation with nested objects and an open-world side effect (a real email), the description covers purpose, consent, recipient handling, privacy limits, sibling routing, and expected reply text. An output schema exists, and the description still shows the canonical reply string, which is more than enough for correct invocation.

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 description coverage is 100% and each field already carries its own constraints, enums, and ranges, so the schema does the heavy lifting. The description reinforces the collection policy ('Ask only for the child's first name') and repeats the note restrictions rather than adding new field-level meaning. Baseline 3 is appropriate for fully documented parameters.

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, channel, sender, and payload: 'Sends one email from SATCoach (hello@satcoach.ai) to a parent, with a link to set up SATCoach for their child', then enumerates the carried data. It also distinguishes itself from both siblings by name (about_satcoach, skill_check), so an agent can route correctly without opening a schema.

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 an explicit gating condition ('only when a parent wants SAT help for their child and wants SATCoach to follow up'), a required pre-call workflow (show the parent exactly what will be sent, get a yes, then set consent to true), and explicit exclusions with the alternative tool for each ('student asks for themselves', 'use about_satcoach', 'use skill_check'). Nothing is left to inference.

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

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

The description goes well past the annotations, which only declare readOnly/idempotent/non-destructive. It explains the whole flow (parent opens the link and hands it to the student, results appear on SATCoach's site and arrive by email) and, critically, that "Nothing about the check or its results comes back to you" — the single most important behavioral fact for an agent deciding how to frame this to a family.

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?

It is front-loaded with purpose and flow, then moves to usage and finally the parameter constraint, with no filler sentences. The only redundancy is the trailing enum restatement, which duplicates what the schema already enforces.

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 one-parameter read-only tool with an output schema and full annotations, nothing an agent needs is missing. The out-of-band result delivery is disclosed precisely, so the agent knows not to expect score data in the response, and the caller's handling instructions (parent gets the link) are stated.

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?

With 100% schema description coverage and a single enum-constrained parameter whose description already says only "SAT" is supported, the schema carries the parameter semantics. The description's closing "test must be \"SAT\"" merely restates the enum, adding no new meaning, so the baseline of 3 applies.

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 opens with a specific verb and resource: "Returns the link and steps for SATCoach's free SAT skill check," followed by concrete scope (20 questions, ~15 minutes, no account). It is immediately distinguishable in practice from about_satcoach and share_family_context, though it never names those siblings 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?

It gives explicit positive triggers ("Use it when a family wants a short SAT diagnostic, or wants to find out where a student stands") and explicit exclusions ("Don't use it to get questions, check answers, or estimate a score in this conversation"). The when-to-use and when-not-to-use conditions are both spelled out with the reason.

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