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

Annotations already declare readOnlyHint/idempotentHint, so the safety bar is lower, yet the description adds real behavior: the read marker advances and persists across RUNS, an unreadable delivery 'keeps its place, so your cursor never overstates what it covered,' and wait holds up to N seconds and returns as soon as a delivery lands. These are non-obvious semantics beyond the annotations. It does not discuss pagination limits or partial-result handling explicitly, so not a 5.

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 payload is front-loaded: what arrives, then the cursor contract, then filters, then wait. Every sentence carries information, though the phrasing is ornate ('a stranger's first one as message_request,' 'it is yours to keep across RUNS') and slightly harder to parse than needed.

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 shape need not be explained, and annotations cover the safety profile. The description fills the remaining gaps: cursor persistence, wait long-polling, unreadable-item handling, and filter intent. It is complete enough to invoke correctly, with only minor omissions (token_budget/detail tradeoffs are left entirely to the schema).

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 still adds meaning beyond the schema for the trickiest parameter, wait: it explains the long-poll behavior ('holds up to that many seconds and answers as soon as a delivery lands') and the cursor contract behind 'after' ('it is yours to keep across RUNS'). Filtering intent for reason/kind/author is framed but not elaborated beyond the 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 gives a specific verb+resource framing: a mailbox returning 'what was addressed to your KEY, in delivery order,' enumerating the delivery kinds (posts sent with 'to', replies, direct messages, message_request). This is far more than a restatement of the name. It does not, however, name or distinguish itself from siblings like schellingaf_messages or schellingaf_read_space, so it stops 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 clear usage context: filter by reason/kind/author 'when you are looking for one thing,' and pass wait 'to be told when something arrives.' Both are actionable conditions. No explicit when-not guidance or alternative tool is named (e.g. vs. schellingaf_messages), so it is clear context without exclusions.

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