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

Annotations already cover readOnlyHint, openWorldHint, idempotentHint, and destructiveHint, so the safety profile is well defined. The description adds that the tool automatically adds 'other' and 'insufficient_context' labels, which is a behavioral trait beyond the annotations. This is useful but minimal; it does not explain output format or edge cases, which is acceptable given the output schema 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?

The description is two sentences with no filler. The core action is front-loaded, and the additional label behavior and usage constraints are stated efficiently. Every sentence contributes meaning without redundancy.

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?

Given the tool's complexity—nested item objects, a labels mapping, instructions, and an output schema—the description is too sparse. It does not explain how to populate the parameters correctly or what the output represents. While the output schema covers return values, the input semantics are largely unexplained, leaving an agent to infer from parameter names alone.

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 schema has 0% description coverage, so the description must compensate. It only hints at 'items' and 'caller-supplied labels' but does not explain that labels is an object mapping label names to descriptions, that items are structured objects with id/text/metadata, or what instructions should contain. This is insufficient for a tool with nested objects and three required parameters.

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 clearly states the action (assign labels to items), the resource (items), and the constraint of exactly one caller-supplied label per item. It also distinguishes itself from prose generation and from the sibling search_and_rank by being a classification tool. The phrase 'using Jev typed decisions' is specific enough to signal a distinct mechanism.

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 explicitly says 'Use for English bulk classification, not prose generation,' which gives positive and negative usage context. However, it does not mention the sibling tool search_and_rank or provide a when-not-to-use alternative beyond prose generation, so guidance is partial.

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, idempotentHint=true, destructiveHint=false, so the safety profile is covered. The description adds behavioral context beyond annotations: it reveals that the tool runs ripgrep internally, that it uses Jev for relevance ranking, and that it returns only relevant matches. It also discloses the 'English judgments only' limitation. This adds meaningful behavioral transparency beyond the structured 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?

Three sentences, each earning its place: the first states the mechanism, the second gives the trigger condition, the third adds a constraint. The most important information (read-only, ripgrep, Jev ranking) is front-loaded. No wasted words.

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 tool has an output schema, so return values are covered. The description explains the core pipeline (ripgrep + Jev), the trigger condition, and the English-only constraint. It doesn't explain what 'relevant' means in terms of the relevance_criterion parameter, but the parameter name and schema make that reasonably clear. For a 6-parameter tool with an output schema, this is nearly complete. The only gap is a bit more detail on how the relevance_criterion interacts with Jev, but that's a minor omission.

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. The description mentions 'patterns' implicitly via 'ripgrep' and 'relevance_criterion' via 'relevance', but it doesn't explain the semantics of mode, top_k, exclude_globs, include_globs, or the relationship between patterns and relevance_criterion. The description adds some context (the tool is a search+rank pipeline) but doesn't fully compensate for the 0% schema coverage. Baseline 3 is appropriate because the description gives a general sense of the parameters but leaves the agent to infer details from 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 states a specific verb ('run read-only ripgrep'), a resource ('current workspace'), and a distinct post-processing step ('use Jev to return only relevant matches'). It also names the sibling tool 'classify_items' implicitly by contrast? Actually it doesn't name the sibling, but it clearly distinguishes itself from a raw search by describing the ranking step. The verb+resource+scope is specific enough to differentiate from a generic search tool.

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 explicitly says 'Use automatically when raw search would exceed 20 candidates or about 4K tokens.' This is a clear when-to-use condition. It also implies when not to use it (when raw search is small enough), and the 'English judgments only' note adds a constraint. This is strong usage guidance.

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