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

Beyond the annotations (readOnlyHint, idempotentHint, destructiveHint), the description adds substantive behavioral context: 'no chain call' (offline behavior not captured by readOnlyHint), 'Read-only, deterministic' (determinism goes beyond idempotency), and 'Needs your API key' (an auth requirement absent from annotations). None of this contradicts the annotations — it complements them. This is exactly the extra context the rubric rewards.

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 dense but every sentence earns its place: operations, usage context, exclusions, return shapes, and behavioral notes. The 'no chain call' scope is front-loaded. It is somewhat long, but given four distinct sub-operations and no output schema, the length is justified. A list format could improve scannability, which prevents a 5.

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 multi-mode tool with 4 parameters and no output schema, the description is thorough: it documents all four operations, their input requirements, and their return shapes (essential since no output schema exists), plus workflow context, exclusions, and behavioral guarantees. Minor gaps remain — the full unit list for convert is vague and error/edge-case behavior (e.g., invalid address handling) is not addressed.

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. The description raises it by adding per-operation parameter semantics the schema lacks: it explains what `op` values do, that `from`/`to` apply 'convert only', that `input` means address/value/signature/text depending on op, and it maps each op to its return fields (checksum → valid_address, checksummed; convert → value, from, to, result). The only minor gap is that valid units are hinted as 'wei…gwei…eth' without full enumeration.

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 'Offline EVM utilities, no chain call' — a specific verb phrase naming the resource and scope. It enumerates all four operations (checksum, convert, selector, keccak) with precise definitions for each, and distinguishes itself from siblings by naming what it is not: 'Not for verifying a signature (use verify_signature) or decoding calldata (use decode_calldata)'. An agent can tell exactly what this tool does and how it differs from rpc_call, verify_signature, and decode_calldata.

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 when-to-use guidance: 'Use it for address hygiene and unit maths before an rpc_call' — a concrete workflow context. It also names alternatives with conditions: 'Not for verifying a signature (use verify_signature) or decoding calldata (use decode_calldata)'. This is the textbook case of explicit when/when-not/alternatives guidance.

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