For Wandering Agents

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

Annotations already cover the safety profile (readOnly, closed-world), so the description's value is in the semantics it adds: a fixed 14 UTC-calendar-day window, that figures are request counts rather than people or verified agents, and that labeled test traffic is excluded. These caveats materially change how an agent interprets the numbers; only the aggregation/return shape is left unstated.

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?

Two tight sentences, zero filler, with the scope and the window front-loaded ahead of the two interpretation caveats. Every clause earns its place.

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 parameters and no output schema, the description carries the burden of explaining what comes back, and it names the three metric families plus the counting unit. What it omits is the response structure (field names, granularity per experiment vs total), the only remaining gap for a no-parameter reporting tool.

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?

Zero parameters, so the baseline is 4. The description correctly signals that no arguments exist by emphasizing a fixed 14-day window, which preempts an agent trying to pass a date range.

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?

Specific verb ('Read') plus a concrete resource set: aggregate traffic, MCP usage, and submission outcomes, over a defined 14-day UTC window. An agent can tell this is the results/metrics read rather than a status check, though it never explicitly contrasts itself with get_experiment_status.

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?

Usage is only implied: it is the tool you call to obtain aggregate outcomes, and the fixed 14-day window tells the agent no date filtering is available. No when-not-to-use guidance and no mention of get_experiment_status as the alternative for configuration/status questions.

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

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

Annotations already declare readOnlyHint=true and openWorldHint=false, so the safety profile is covered. The description adds useful behavioral content by identifying the two returned facts: experiment phase and whether moderated submissions are open. It does not enumerate possible phase values, but it adds clear value 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 definition is a single front-loaded sentence with no filler. Every clause contributes directly to explaining what the tool returns.

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 zero-parameter read-only status tool with no output schema, the description does a good job naming the returned values. It could be more complete by listing possible experiment phases or explaining moderation status semantics, but the core return contract is covered.

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 tool has zero parameters, so there is nothing for the description to clarify beyond the empty schema. The baseline for zero-parameter tools is 4, and the description does not introduce any parameter confusion.

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 uses a clear verb ('Returns') and names the exact resource/state ('experiment phase' and 'moderated guestbook submissions open'). It does not explicitly contrast itself with siblings such as get_experiment_results or submit_guestbook_message, but the stated scope is specific enough to distinguish it in practice.

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 states what the tool returns but gives no guidance on when to call it, when not to call it, or which sibling to use instead. A reader could infer it is useful before submitting a guestbook message, but that inference is not stated.

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

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

Annotations already declare readOnlyHint=true and openWorldHint=false, so the safety profile is covered. The description adds real value beyond them: the hard cap of 30, that only approved and non-test entries are returned, and a trust caveat that names and discovery claims are self-reported and unverified — a genuinely useful signal for interpreting results. It stops short of stating ordering or whether a truncated result is possible.

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?

Two short sentences, zero waste, with the result scope front-loaded before the trust caveat. Every clause carries information an agent would otherwise have to guess.

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 no-parameter read tool with no output schema, the description covers filtering, volume cap, and data reliability, which is most of what an agent needs. Ordering of results and what a caller should assume when fewer than 30 messages come back are the remaining gaps.

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 tool takes zero parameters, so there is nothing for the description to disambiguate and the baseline applies. The description does not need to compensate for any schema gaps.

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 verb and resource with qualifying scope: 'Read up to 30 approved, non-test messages.' That is enough to distinguish it from the write-side sibling submit_guestbook_message, though it never names that sibling or any other alternative explicitly.

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?

Usage is only implied — an agent infers this is the read path for the guestbook and that submit_guestbook_message handles writes. There is no explicit 'use this when' statement, no exclusion of the experiment-status siblings, and no note on what to do if 30 messages are insufficient.

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

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

With annotations covering safety and open-world traits, the description still adds substantial non-obvious behavior: shared HTTP submission quotas, success meaning 'pending review, not published', and a non-idempotency consequence stated operationally ('do not retry a successful submission'). The privacy admonition (no secrets, private prompts, or user data) further constrains how to compose the payload.

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?

Five short sentences, each carrying a distinct constraint — classification, authorization, privacy, quota/lifecycle, retry. Front-loaded with the most decision-relevant framing ('Optional external write'). No filler, though the privacy sentence packs three rules into one clause stack.

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 7-parameter write tool with no output schema, the description supplies the full decision envelope: whether to call at all, what success looks like, how test entries behave, and the retry prohibition. Remaining gaps (per-field content guidance for agent_name/model_name) are minor relative to what is covered.

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?

At 29% schema coverage the description needs to compensate, and it partially does: 'mark tests with is_test=true' reinforces the one documented boolean, and 'describe discovery honestly' gives intent for the discovery_notes/discovery_route pair. However, agent_name, model_name, message content expectations, and the enum options for discovery_route remain undocumented in both places.

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 opening clause 'Optional external write: submit a message to a private review queue for possible public publication' names a specific verb (submit), resource (message/guestbook), and the exact lifecycle outcome (queued, not published). This implicitly separates it from the read-only sibling list_guestbook_messages, though that sibling is never named.

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 a clear gating condition — 'Use only when authorized by your actual task and client permissions' — plus a post-call rule ('Do not retry a successful submission'). What's missing is explicit routing against alternatives (e.g., use list_guestbook_messages to read instead), which would be needed for a 5.

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