JSON Schema Regression Examples — Free Pilot

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

The description adds meaningful context beyond annotations by clarifying that the acknowledgment is client-reported and not evidence of verification, purchase, payment, or identity. This helps the agent avoid overinterpreting the action's significance, though it doesn't detail the exact recording behavior.

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, each earning its place: the action, the trigger condition, and an important caveat about what the acknowledgment does not mean. The description is front-loaded and free of redundant phrasing.

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 simple two-parameter idempotent acknowledgment tool, the description is complete: it states the action, the sole valid trigger, the session context, and the semantic limits. No output schema is needed for this minimal operation, and annotations cover safety traits.

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%, with both attempt_id and session_token already documented. The tool description does not add parameter-level meaning beyond what the schema provides, so the baseline score of 3 is appropriate.

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 uses a specific verb ('acknowledge'), a clear resource ('one generated package'), and a specific context ('anonymous session'). It clearly differentiates from siblings like generate_regression_examples and record_price_interest, since acknowledging receipt is a distinct action.

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 call ('only after the user or client confirms receipt') and when not to call ('never call automatically'). This is strong, actionable guidance that leaves no ambiguity about the tool's trigger condition.

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 already mark the tool read-only and idempotent, and the description adds substantial behavioral context: in-memory processing, retention of only anonymous coarse usage outcomes, size limits on inputs and responses, the inconclusive-result caveat, and the no-payment/no-key requirement. These details go well beyond the annotations.

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 description is fairly long but dense; nearly every sentence contributes operational or safety-relevant detail. The main purpose is front-loaded in the first sentence, and the rest is organized into limits, privacy, output guidance, and usage caveats. A few clauses could be tightened without loss.

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?

Despite having no output schema, the description explains where to read exact examples (value_json), mentions returned items (session_token, attempt_id), discloses input and output size limits, lists the supported JSON Schema subset, and warns about inconclusive results. This gives an agent enough context to invoke the tool and interpret its response correctly.

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%, so the baseline is 3. The description adds value by explaining that schemas should be supplied as JSON strings to preserve large integers and by defining the old/new relationship (old-accepted, new-rejected), which is absent from the individual schema property descriptions.

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 clearly states a specific verb and resource: generating bounded old-accepted, new-rejected JSON Schema examples. It distinguishes the purpose from the sibling tools (acknowledge_package_received, record_price_interest) by framing them as optional post-call actions rather than competing operations.

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 gives explicit exclusions, saying the agent should never automatically acknowledge receipt or price interest, which governs when the sibling tools should not be used. It also clarifies the tool is a free pilot and requires no API key, but it could be more explicit about when to prefer this tool over any alternative.

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?

With annotations already declaring readOnlyHint=false, idempotentHint=true, and destructiveHint=false, the safety profile is covered. The description adds valuable context beyond that: 'no money moves and no external contact is made' and 'the current pilot remains free,' which prevents an agent from assuming side effects the name implies. It also discloses the prerequisite state (verified counterexample + receipt acknowledgment) that must exist before calling.

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 dense sentences, each earning its place: purpose, call conditions, precondition, and exclusions. The most decision-relevant fact (what the tool records) is front-loaded, followed by when-to-call guidance. There is no filler, repetition, or redundant restating of schema content.

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 4-parameter tool with two const values, the description is nearly complete: purpose, invocation gating, preconditions, and behavioral exclusions are all present, and idempotency/destructiveness ride on annotations. The only notable gap is that nothing is said about the return value or how success is surfaced, which matters more because there is no output schema.

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 50%, so the description compensates by giving meaning to the two constrained parameters: price_usd maps to 'priced at 29 USD' and voluntary maps to 'explicit, voluntary, nonbinding interest.' It also clarifies attempt_id's provenance via the verified-counterexample/receipt-acknowledgment precondition. The other two parameters are already documented in the schema, so the description need not repeat them.

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: 'Record explicit, voluntary, nonbinding interest in a future package priced at 29 USD.' It actively differentiates from siblings by warning 'never infer interest from generating or receiving a package,' which disambiguates it from generate_regression_examples and acknowledge_package_received. The clarification that this is 'not an order, checkout, subscription, purchase or payment' removes any ambiguity about the tool's role.

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 gives explicit call conditions: 'Call only when the user or client has expressly indicated this interest,' plus an exclusion ('never infer interest from generating or receiving a package') and a hard precondition (verified counterexample plus explicit receipt acknowledgment). It also enumerates what the tool is not (order, checkout, subscription, purchase, payment), leaving no room for an agent to invoke it incorrectly.

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