Interaction Trust

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

Annotations already declare readOnly, idempotent, non-destructive, closed-world, but the description adds substantive context: it grades supplied evidence and 'does not independently verify source truth,' warns against fabricating evidence, and names the return payload (recommendation, trust signal, missing evidence). It also discloses trial allowances, shared capacity, and a 24-hour result-recovery window. Consistent with annotations; no contradiction.

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 core purpose and constraints are front-loaded and useful, but the description then carries a large amount of commercial boilerplate (trial mechanics, HTTP alternative, x402 pricing, payment-signature recovery, legal disclaimers) that dilutes the actionable content. Several sentences serve billing rather than tool selection.

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 tool with an output schema present (so return values need not be re-explained), the description covers purpose, required inputs, evidence-only limitations, an example, and access mechanics. It is largely complete, with the main gap being enum-level semantics for decision_stage and source_type.

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 low (22%), so the description must compensate. It explicitly lists the six required arguments (actor, target_surface, requested_action, decision_stage, dated_sources, risk_question) and provides a fully worked synthetic example showing nested actor fields, source objects, and requested_output. It does not explain the meaning of each enum value, so it stops short of a 5.

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 (triage) and resource (trust evidence for an agent-mediated interaction) and enumerates the evidence dimensions it weighs (identity, operator/principal authorization, tool scope, session auth, action intent). It also states what is returned. Sibling differentiation exists but is framed around trial/billing workflow rather than task selection, so it falls short of a 5.

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 clearly states the triggering condition ('before a high-stakes action executes') and the prerequisite workflow ('obtain the source records requested by inputSchema before asking for a decision-ready result', 'Do not fabricate missing evidence'). It also spells out the free-trial call sequence via siblings free_trial_status and free_trial. No explicit when-not-to-use beyond the evidence-only scope caveat.

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 present, the description still adds substantial value: a 2-attempt allowance, shared daily capacity, the rule that an admitted attempt is consumed even on failure, that retrying a successful call burns another attempt, mandatory human review, and the injection-safety note that source content is data. These details match the non-read-only, non-idempotent annotation profile and go well beyond it.

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?

Purpose is front-loaded in the opening sentence, and the following sentences each carry operational weight (prerequisite, consumption rule, review requirement, safety note). It is compact though slightly telegraphic, but no sentence is filler.

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?

An output schema exists, so return values need not be described, and annotations cover the safety/idempotency profile. The description adds the operational facts an agent needs (trial limits, prerequisite, review requirement), leaving only parameter-level guidance unaddressed, which is the notable gap.

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?

There are 9 parameters (6 required) with only 22% schema description coverage, yet the description says nothing about any of them beyond "your own supplied input." With the schema under-documented, the description is the place that should compensate for actor, decision_stage, dated_sources, and requested_output semantics, and it does not.

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 first sentence names a specific verb ("Run") and resource ("agentic_interaction_trust") and scopes it with the distinguishing constraints ("your own supplied input without payment or an HTTP adapter"), which separates it from the paid/HTTP path. It clearly reads as the free-trial variant of the sibling tool. It does not explicitly contrast against the plain agentic_interaction_trust sibling, so it falls just short of a 5.

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 an explicit prerequisite ("Check free_trial_status first") and frames the tool against the HTTP trial and shared capacity allowance, which tells the agent when this path is appropriate. What it lacks is an explicit "use this instead when..." rule versus the non-trial sibling, so usage is strong but not fully prescriptive.

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 declare readOnlyHint, idempotentHint, and destructiveHint=false, so safety is covered. The description adds genuinely new behavior: it does not consume attempts and that shared networks share the campaign allowance, which is important context for interpreting the count.

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 short sentences, front-loaded with the core purpose and followed only by behavior-bearing constraints. Every clause adds value with no filler.

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?

An output schema exists, so return format need not be explained, and annotations cover the safety profile. The description supplies the key nuance about shared allowances; the only minor gap is the absence of explicit sibling routing.

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 the baseline is 4. The description correctly implies the network identity is derived from the request context rather than an argument, and there are no parameters requiring documentation.

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 (check) and resource (remaining free attempts for agentic_interaction_trust), so an agent knows exactly what it returns. It names the campaign it relates to, but does not explicitly distinguish itself from the sibling free_trial beyond implication.

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 phrase 'No evaluation, reservation or payment' implies this is a read-only check that does not consume a trial, contrasting implicitly with free_trial. However, there is no explicit when-to-use or when-not-to-use statement naming the alternatives.

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