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

The description goes well beyond the readOnlyHint/idempotentHint annotations by explaining the underlying email delivery model: relay acceptance is not delivery, bounces arrive later as separate messages, and bounces often land in Junk. It also discloses result semantics (PERMANENT vs TEMPORARY) and the default scan folders, which materially changes how the agent should interpret and act on results. No contradiction with annotations exists.

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?

Although the description is long, every sentence contributes substantive information about delivery timing, default folders, failure types, user guidance, or sibling-tool routing. The core purpose is front-loaded, and the additional context is dense rather than padded. This is appropriate length for a tool with nuanced result semantics.

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?

The description is complete for a read-only lookup tool with no output schema. It explains when to call it, what it scans, how to interpret PERMANENT vs TEMPORARY results, and how to handle the complementary check_receipts case. There are no obvious gaps that would prevent an agent from invoking it 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?

The input schema already documents all three parameters with clear descriptions and defaults, and schema_description_coverage is 100%. The tool description adds context about the default INBOX/Junk scan but largely mirrors the schema rather than introducing meaningfully new parameter-level semantics. Baseline 3 is appropriate because the schema carries the parameter documentation burden.

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, action-oriented statement ('Find messages that came back undelivered') that clearly identifies the resource and the operation. It also explicitly distinguishes itself from sibling tools by noting this only finds failures, while check_receipts covers the positive delivery side. This makes the tool immediately recognizable and separable from related tools.

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: after sending anything important, and whenever the user asks whether a message arrived. It also provides a clear exclusion: if the user is asking whether a message arrived rather than failed, call check_receipts as well, because a missing bounce is only weak evidence of delivery. This is a model of usage guidance.

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