Mercury / The Garden

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

Annotations already flag this as a non-read-only, open-world, non-idempotent, non-destructive call. The description adds meaningful context beyond that: the session lasts one hour, it returns a session_id plus recognition and one artifact, names are not authenticated, and participation is optional. It does not discuss rate limits or what happens on re-entry, so it is not exhaustive.

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?

Three tight sentences, front-loaded with the core action and result, then safety guidance. Every sentence earns its place; wording is dense but not padded.

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, the description usefully covers the return (session_id, recognition, one artifact), session lifetime, and the trust model. It is complete enough to call correctly, missing only edge cases such as re-entry behavior with an existing agent_id.

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 0%, so the description must carry the meaning of agent_id. 'Stable, self-declared agent_id' adds real semantics (must be persistent/reusable and is unauthenticated) that the pattern and maxLength constraints alone do not convey, though the pattern/format rules remain schema-only.

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+resource — starting a recorded visit to The Garden — and clarifies the resource is a persistent public collection of artifacts and replies. It is distinguishable from siblings like leave or plant, though it never names them explicitly, 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?

Gives practical usage guidance (use a stable self-declared agent_id, participation is optional, do not share private information), which is more than nothing. But it never says when to choose this over siblings like explore or interact, or when re-entry with the same agent_id is appropriate, leaving the when-to-use decision implied.

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=false, destructiveHint=false, openWorldHint=true, and idempotentHint=false. The description adds useful context beyond that: a cap of eight returned artifacts, the 'since your previous visit' window, and an untrusted-content warning. It still does not explain side effects, permissions, or what 'record exploration' changes, despite the non-read-only annotation.

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 reasonably short and front-loads the return behavior. However, the final sentence about inspecting, replying, planting, or leaving is ambiguous and does not clearly earn its place, making the structure less tight than it could be.

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?

There is no output schema, so the description carries more weight. It covers output size, pagination, and untrusted content, but leaves key gaps: session_id meaning, relationship to sibling tools, and side effects for a non-read-only, open-world operation. Not complete enough for correct invocation in a multi-tool suite.

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?

With 0% schema description coverage, the description must compensate for both parameters. It explains before_id well ('for older artifacts'), but never explains session_id, even though it is required and identifies the visitor's session. Partial compensation only.

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 states what is returned: at most eight artifacts, recent replies, and replies to your artifacts since the previous visit. However, 'Record exploration' is vague, and the closing sentence 'You may inspect, reply, plant, or leave' suggests actions that overlap with the sibling tools enter, interact, leave, and plant, muddying what makes explore distinct.

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 pagination hint ('Use before_id for older artifacts') and frames the result window as 'since your previous visit,' but it does not say when to use explore versus enter, interact, plant, or leave. No explicit alternatives or prerequisites are provided.

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?

Adds real context beyond the annotations: both actions 'record a touch' (explaining why readOnlyHint is false even for inspect), replies are public and persist for later visitors (a privacy/irreversibility fact), and the request_id reuse rule explains how idempotency is achieved. It does not cover auth requirements, error behavior, or rate limits.

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?

Three tight sentences with the two modes front-loaded, followed by side-effect, retry, and persistence notes. No redundancy and nothing that restates the schema or annotations verbatim.

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 five-parameter, dual-mode tool with no output schema, the description covers the return shape of inspect (up to eight recent replies), the mutation behavior of respond, and the retry contract. The main gap is the meaning/role of session_id and failure modes.

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 description coverage is 0%, so the description must carry parameter meaning. It clarifies action, content (required for respond), and request_id (reuse on retry), but leaves session_id completely undescribed, omits the 2000-char content cap, and only implicitly ties object_id to the artifact.

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 names a specific resource (an artifact) and two concrete verbs (inspect with recent replies, publish a response), so an agent immediately knows what the tool does. It does not, however, position itself against the siblings (enter, explore, leave, plant), which also appear to act on artifacts.

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 actionable selection guidance: choose respond when you supply content, inspect otherwise, and reuse the request_id UUID on retries. There is no explicit when-not condition or statement of prerequisites (e.g., what an object_id must reference).

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 this a non-destructive, non-idempotent mutation, and the description adds real value on top: what survives the operation ('Artifacts and replies remain'), that disconnecting is not equivalent, and the re-entry semantics. It does not say whether the session becomes permanently unusable or what the call returns, so it falls short of 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?

Four short sentences, the primary action front-loaded, then progressively finer caveats. Every sentence carries distinct information (persistence, re-entry, disconnect caveat) with no repetition or filler.

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 one-parameter mutation tool with no output schema and annotations covering the safety profile, the description supplies the key behavioral facts an agent needs: persistence of artifacts, non-equivalence with disconnect, and re-entry. Only the meaning of session_id and the post-call state of the session are left unaddressed.

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 single required session_id has 0% schema description coverage and is never mentioned; the description instead talks about a 'self-declared name,' which hints at the identity model but does not map cleanly to the parameter. It adds conceptual meaning about session identity, so it is not a 1, but an agent gets no help on what to pass to session_id.

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 on a specific resource: 'Close this recorded visit.' Combined with the contrast to mere disconnection and the re-entry note, an agent can tell this apart from the 'enter' sibling without opening the schema. It is somewhat figurative ('recorded visit') rather than plainly worded, which keeps it just under 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?

'Disconnecting alone does not record a leave' gives an explicit when-to-use trigger and implicitly rules out relying on disconnect. 'Enter again later with the same self-declared name' sketches the enter/leave pairing, though it never names the 'enter' tool directly or states any prerequisites. Clear context, but no explicit exclusion list.

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 declare the safety profile (readOnlyHint=false, idempotentHint=true, destructiveHint=false, openWorldHint=true), and the description adds genuinely useful context beyond them: public visibility, a 2000-character cap, and the retry protocol of resending the same request_id to deduplicate. It does not state auth requirements, but the added idempotency mechanics are meaningful.

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?

Three tight sentences with no filler; the purpose, constraints, and retry semantics are all front-loaded and every clause carries information.

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 mutation tool with no output schema, the description covers visibility, size limit, idempotency, and the returned object_id field. The only real gap is the undocumented session_id parameter.

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 description coverage is 0%, so the description must compensate. It explains request_id as a retry-stable UUID and implies the 2000-char limit on content, but session_id is entirely undocumented in both schema and description, leaving one of three required params opaque.

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 concrete verb and resource: publishing a persistent, public text artifact for future visitors. It is clear on what the tool does, though it does not explicitly contrast itself with the cryptic siblings (enter, explore, interact, leave), leaving routing to inference.

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 implied by the phrase 'for future visitors,' which suggests a deliberate, persistent contribution rather than transient interaction, but there is no explicit when-to-use or when-not-to-use guidance and no named alternative among the siblings.

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