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

Annotations declare readOnlyHint=false and openWorldHint=true, and the description adds the crucial disclosure that the default call may modify workspace files, with "plan" as the safe alternative. It also exposes statefulness (conversationId continues a prior run), a trait the annotations do not capture. Missing details on timeouts and what happens under sandbox keep it below 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?

Three short sentences, front-loaded with the action, then the safety-relevant default and the override, then the state-continuation tip. Every sentence earns its place with no 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 10-parameter wrapper with no output schema, the description covers the two things an agent most needs to know before calling: that it can mutate the workspace and how to continue a conversation. It does not describe the shape of the returned answer or partial-output behavior on timeout, a minor gap given full schema coverage.

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 100%, so all ten parameters including the mode enum and conversationId are already documented in the schema. The description only reinforces defaults and the continuation pattern, adding marginal meaning beyond structured fields; baseline 3 applies.

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 concrete verb+resource: hand a task or question to the Antigravity CLI and return its answer. That is specific enough to distinguish it from antigravity_web_search or antigravity_shell, though it never explicitly frames itself as the general-purpose delegation tool relative to its specialized siblings.

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 real context for the mode parameter (default may edit files, use "plan" for a read-only answer) and for conversationId reuse, which is when-to-use guidance for a call pattern. It stops short of routing guidance against the siblings (e.g., use antigravity_web_search for lookups instead), so the agent must infer tool selection.

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 'Read-only' simply restates structured data and earns no credit. The description adds no behavioral detail beyond that: nothing about how multiple filePaths are handled (one question across all files?), whether partial results are common, or what the model default is when 'model' is omitted.

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, purpose front-loaded, no padding. The trailing 'Read-only' clause is the one line that does not earn its place because it duplicates the annotation.

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 100% schema coverage and an annotation-declared safety profile, and no output schema to explain, the description covers what an agent needs to invoke it. Gaps are secondary: how multiple files are combined and how the 'model' parameter interacts with the default.

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 100%, so every parameter including model, prompt, filePaths, and timeoutSec is already documented with examples and constraints. The description contributes nothing beyond the schema, making the baseline 3 the correct ceiling.

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 pair: read images, screenshots or PDFs and answer a question about them. An agent can tell it is a media Q&A tool rather than a shell, search, or model-listing tool. It stops short of naming any sibling explicitly, so differentiation is inferred from the resource types rather than stated.

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?

There is no when-to-use or when-not-to-use guidance, and no alternatives are named despite five siblings in the same family. The only implied condition is 'you have an image or PDF and a question', which is restated from the purpose rather than guidance for choosing this tool.

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 'Read-only' merely repeats structured data. The one genuinely additive detail is that the tool both generates ideas and selects the most promising ones – a two-phase behavior not visible in annotations. It still omits what the selection criteria are and what the returned result looks like, so it goes modestly beyond the annotations rather than deeply.

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 with no filler, and the core action is front-loaded before the safety note. Every clause contributes something an agent can use.

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 is the only place an agent could learn what comes back – and it says only that promising ideas are 'picked', without format, count, or ranking detail. For a 6-parameter tool with no output schema, that leaves a real gap. Parameters and safety are otherwise adequately covered, so it is not a severe failure, but it is incomplete.

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 100%, so every parameter (ideas, model, topic, method, context, timeoutSec) is already documented in the schema, including the method enum values and the default of 8 ideas. The description adds no syntax, format, or interaction detail beyond that baseline. A 3 is the correct floor when the schema does all the work.

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?

Names a specific verb and resource: generate ideas on a topic using a chosen method, then pick the most promising ones. That two-stage purpose is clear enough to distinguish it from a plain chat tool. It does not, however, contrast itself with siblings like the base `antigravity` tool, leaving the boundary implicit.

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 offers no when-to-use guidance, no exclusions, and no alternatives. 'Read-only' hints at the safety profile but says nothing about when brainstorming is the right tool versus `antigravity` or `antigravity_web_search`. The schema does point at `antigravity_models` for model ids, but the description itself carries no routing advice.

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 safety and locality are covered. The description adds the useful constraint that results are scoped to the caller's plan, but says nothing about result shape, ordering, or whether the list can change over time. Incremental value over annotations is modest.

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 with no filler. The purpose is front-loaded and the chaining instruction follows immediately; 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?

There is no output schema, so the description carries the burden of describing returns, and it only vaguely does ('list the models'). For a tiny discovery tool the gap is small, but an agent still lacks any sense of the returned fields (id, name, capabilities) or list size.

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 are no parameter semantics to explain; the baseline for a 0-param tool is 4. Mentioning `model` refers to the sibling tool's parameter, not this one, so no credit or penalty either way.

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: listing models available to the Antigravity plan. It also implicitly distinguishes itself from the sibling `antigravity` tool by naming that tool as the consumer of the ids returned. The scope word 'your Antigravity plan' is a little fuzzy but the intent is unambiguous.

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 second sentence gives concrete downstream usage: take an id from this output and pass it as `model` to the `antigravity` tool. That is real routing guidance. It stops short of stating when not to call this (e.g., when the model id is already known), so it is not a full 5.

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. The description adds genuine value beyond them: it discloses the quota-free nature of the call and that the check actually attempts to start the agy binary, which is more than a trivial network ping.

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, front-loaded with the purpose and followed by the quota note. No filler; 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?

For a zero-required-param health check with no output schema, the description covers what is checked and the cost profile. It omits any indication of what a pass/fail response looks like, which is a minor gap rather than a blocking one.

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 100%, so the lone 'message' parameter is documented there as text to echo back. The description does not mention this parameter or reconcile it with the health-check purpose, so it adds nothing over the schema — baseline 3.

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 target: verifies the server is up and the agy binary can be started. This is a health-check distinct from all siblings (search, shell, brainstorm, models), so an agent can select it without inspecting the schema.

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 (connectivity/preflight check) and 'Does not use any quota' hints it is safe to call freely, but there is no explicit when-to-use statement or comparison to sibling tools that might also report server health.

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 destructiveHint=true and openWorldHint=true, so the risk profile is partly covered. The description still adds real behavioral value beyond them: the required permission, and the specific failure mode (permission-denied error naming the rule to add). It does not elaborate on what executing arbitrary shell commands can affect.

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 sentences, front-loaded with the core capability and mode switch, then the prerequisite. No filler; every clause carries actionable 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 destructive, open-world shell tool with no output schema, the description covers the key operational facts: the mode switch, the permission prerequisite, and the error behavior. It does not address execution environment or side-effect scope, which would round it out fully.

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 100%, so every parameter including `execute`, `cwd`, `model`, and `timeoutSec` is already documented in the schema. The description restates the `execute` semantics but adds no syntax, defaults, or constraints beyond what the schema provides.

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 pair (write or run shell commands) tied to an explicit mode switch (`execute is true`), with the resource being shell commands generated by Antigravity. It is clearly distinguishable from siblings like antigravity_web_search, antigravity_analyze_media, or antigravity_models.

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?

Explains the two usage modes (draft-only vs. run) and names the activation condition for the run mode. It also supplies the prerequisite (`command` permission in agy's settings.json). It does not compare against sibling tools such as plain `antigravity` for non-shell tasks, so it stops short of full when-not guidance.

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 cover readOnlyHint and openWorldHint, but the description goes further: it names the exact permission dependency, the failure mode and its error text, and notes that partial output is returned on timeout. That is concrete operational context an agent cannot get from the structured fields.

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?

Purpose is front-loaded in the first sentence, followed by behavioral caveats, with no filler. "Read-only" is a minor duplication of the readOnlyHint annotation, the only slightly redundant fragment.

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 simple three-parameter read tool with no output schema, the description covers purpose, the permission prerequisite, and failure behavior. The remaining gap is routing guidance versus sibling tools, which is a usage rather than completeness concern.

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 100%, so all three parameters (model, query, timeoutSec) are already self-documented, including the model-list cross-reference. The description adds no parameter-level detail beyond the schema, so the baseline of 3 applies.

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+resource ("look something up on the web and answer with source URLs") and states the return shape, which is more than the tautological name would suggest. It does not explicitly distinguish itself from siblings like antigravity_shell or antigravity_brainstorm, so an agent must infer the boundary.

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 supplies a real usage prerequisite (the read_url permission must be allowed in settings.json, otherwise a permission-denied error naming the rule appears), which is genuinely actionable. However it never says when to prefer this over alternatives such as antigravity_shell for fetching pages, leaving the selection implied.

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