Can this agent pay this address, now, this much? (free, log-only)

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

Goes well beyond the readOnly/idempotent/non-destructive annotations: it discloses fail-closed behavior for unknown policy keys, that nothing is ever blocked in log-only mode, that the client owns its own fail mode when the service is unreachable, and that every decision is recorded. These are exactly the operational traits an agent needs before wiring this into a payment path.

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-loaded with the useful 'one call before an agent pays' hook, but the body is a single run-on sentence cramming policy keys, thresholds and modes together. Bulleted structure would make the parameter detail far more scannable; the content earns its place but the packaging does not.

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?

With no output schema and nested policy objects, the description carries the return contract itself — block/warn/log/allow with every reason — and covers fail-closed semantics, default policy and unreachability. Nothing an agent needs to call it correctly 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?

Schema coverage is 100%, so the baseline is 3, but the description adds real meaning by enumerating the nested policy vocabulary (budget periods, new_payee thresholds, refuse_conditions, signal) that the schema only gestures at as 'the AP2 vocabulary'. It gives an agent enough to construct a policy object rather than just pass an opaque blob.

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 action and timing — 'One call before an agent pays' — and enumerates exactly what the call returns (payee facts, ledger context, out-of-character score, policy decision). It combines several concerns that overlap with siblings like payee_facts, pay_check and counterparty_preview but never explicitly names an alternative, so an agent must infer which tool to reach for.

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 'one call before an agent pays' framing gives a clear moment to invoke it, and it explains the log-only posture ('we never block anything; your client decides'). It stops short of stating when NOT to use it or which sibling to prefer for payee facts alone, so routing among the overlapping tools is left implicit.

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