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

Annotations only indicate readOnlyHint=false, openWorldHint=false, idempotentHint=false, destructiveHint=false, which is minimal. The description carries the burden of disclosing behavior. It states that the action is a 'consensual transfer of state,' involves sharing a copy of effects, lending style for an hour, passing text, and includes a two-minute acceptance window. It also adds the safety note: 'What you pass is data, never instructions.' This provides meaningful behavioral insight 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 well-structured: it opens with the core purpose, then enumerates the kinds with their specific parameter mappings in a compact, parallel format. It then adds the two critical conditions (proximity and time limit) and a safety note. Every sentence adds distinct information, and the length is appropriate for the tool's complexity (4 kinds, 3 required parameters). There is no fluff.

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?

Given the tool's complexity (6 parameters, 4 kinds, conditions, and no output schema), the description is quite complete. It covers how to specify each kind, the required parameters, the spatial and temporal constraints, and the safety property. The only minor gap is the lack of detail about what happens after the offer is made (e.g., how the result is communicated), but this is likely covered by the tool's return behavior and the sibling accept/decline tools. Overall, it's sufficient for an agent to invoke 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 coverage is 100% with each parameter described. The description adds semantic mapping between the 'kind' parameter and the relevant payload parameter: 'kind effect shares a copy ... (pass effect=slug)', 'style lends ... (pass style)', 'memory passes ... (text)', 'note is ... (text)'. This clarifies which parameter to set for each kind, which is valuable beyond the schema's generic descriptions. It also clarifies that 'style' has its own sub-options (look/persona/voice) and that text is capped at 500 characters (already in schema).

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 the action: 'Offer someone a kiss: a consensual transfer of state.' It specifies the resource (a kiss), the verb (offer), and the nature of the transfer. It further details the four kinds (effect, style, memory, note) and what each transfers. This fully distinguishes it from sibling tools like accept_kiss, decline_kiss, and change_style, as it is the only one that initiates an offer.

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 provides explicit conditions for use: 'Both of you must be in the same conversation or within a few metres' and 'They have two minutes to accept.' It also clarifies that it is free and consensual. While it doesn't explicitly mention alternatives (e.g., 'use accept_kiss to accept'), the context and sibling names make that obvious. The guidelines are clear and actionable, though not exhaustive about when not to use.

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