classifier.dev

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

Annotations already declare readOnlyHint=true, idempotentHint=true, and destructiveHint=false, so the safety profile is covered. The description adds meaningful behavioral context: independent per-label scoring, the 0.7 keep threshold, and the 'possibly none' outcome. It doesn't describe the exact output shape, but an output schema exists, so that burden is reduced. Minor gap: no mention of how instructions interact with scoring, but overall strong.

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 sentences, all information-dense and front-loaded. The first sentence establishes the core behavior and contrast with the sibling; the second gives use cases; the third and fourth give operational parameters. No filler or repetition of schema content.

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 multi-label classification tool with a rich schema (100% coverage), an output schema, and safety annotations, the description covers the essential decision points: when to use it, how it differs from classify_texts, how scoring works, and how to cap results. Nothing an agent needs to invoke it correctly is missing.

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 schema already documents all four parameters. The description adds value by explaining the threshold behavior (>= 0.7 kept) and the 'possibly none' outcome, which clarifies the semantics of max_labels and labels. It doesn't add much beyond the schema for instructions, but the baseline is 3 and the description does compensate with useful behavioral detail.

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 ('classify'/'tag'), a resource ('texts'), and the key differentiator: every label that applies, with independent 0-1 scores. It explicitly contrasts with classify_texts and gives concrete use cases (topics of an article, components touched by a ticket), so an agent can distinguish it from siblings without opening 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?

The description explicitly says 'Like classify_texts, but...' and explains when to use this tool ('Use this for tagging... where one answer is not enough'). It also gives practical guidance on setting max_labels and the 0.7 threshold. This is clear when-to-use guidance with an explicit alternative named.

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 mark the tool as read-only and non-destructive, and the description adds genuinely useful behavioral detail: confidence is calibrated with concrete thresholds, answers below 0.5 are characterized, and the 'smart' tier re-asks uncertain items via a reasoning model. This goes well beyond what annotations alone convey.

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 three purposeful sentences: core behavior, use cases, and confidence/tier behavior. Nothing is redundant, and the most important constraint ('exactly one label', 1,000 max) is front-loaded.

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 classification tool with an output schema and read-only/idempotent annotations, the description covers limits, calibration, tier behavior, and actionable guidance for low-confidence results. An agent has everything needed to invoke it correctly and interpret the consequences.

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 burden on the description is lower, but it still adds meaning beyond the schema by explaining what 'calibrated confidence' means and how the 'smart' tier behaves relative to confidence thresholds. This helps the agent reason about parameter choices without opening 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 opens with a specific verb and resource: 'Sort up to 1,000 texts into exactly one of your own labels each, with a calibrated confidence per answer.' This clearly distinguishes it from the sibling classify_multi_label by emphasizing single-label output, and the 1,000-item constraint adds precision.

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 explicit when-to-use context ('many items to triage, route, filter or bucket and do not want to read them all') with concrete examples, and an explicit when-not-to-use rule ('Do not use it for fewer than about five items you can already see — just decide'). This is strong practical guidance for an agent choosing between tools.

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 read-only, idempotent, and non-destructive behavior. The description adds important behavioral traits: it returns only a histogram, not individual classifications, and it counts uncertain items via the confidence threshold. This complements rather than repeats 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 description is compact and front-loaded, stating the core behavior in the first sentence and then giving concrete examples. Every sentence earns its place by defining purpose, usage, or a sibling distinction.

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 rich annotations, full parameter schema coverage, and available output schema, the description supplies everything an agent needs to select and call the tool correctly: scope, scale, result shape, and sibling distinction.

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 parameters are already well documented and the baseline is 3. The description adds no new parameter-level meaning; it only alludes to the output-level concept of uncertainty.

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 opens with a specific verb and resource: classify up to 1,000 texts and return a histogram of label counts. It also distinguishes itself from classify_texts by emphasizing aggregate counts rather than per-item answers.

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 states when to use this tool ('when you want the shape of a corpus') and when not to ('Use classify_texts when you need the answer per item'). This gives the agent a clear routing rule versus the closest sibling.

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, idempotentHint, destructiveHint, covering the safety profile. The description adds valuable behavioral context: threshold default (0.7), returning the two most likely labels, and returning an index for mapping. It also explains the practical meaning of low confidence. This goes beyond annotations without contradicting them.

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 with zero fluff. The core action and return behavior are front-loaded, followed by usage context and mapping detail. Every sentence earns its place, and the structure guides the agent from what it does to why it exists.

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 moderate complexity (4 parameters, 2 required, output schema present), the description covers all critical aspects: threshold behavior, return contents (two labels, index), and how to interpret results. An agent has everything necessary to call it correctly without additional inference.

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 schema already documents all parameters, defaults, and constraints. The description adds no additional parameter semantics beyond what the schema provides, such as clarifying the 'below' threshold's role or the 'labels' list's purpose. It meets the baseline of 3 for high coverage but does not elevate further.

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 and resource: 'Classify up to 1,000 texts and return only the ones whose confidence fell under a threshold'. It clearly distinguishes from siblings by framing this as a post-hoc review tool after bulk classification, which differentiates it from classify_texts or classify_multi_label. The purpose is unambiguous and actionable.

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 gives context: 'Use this after a bulk classification to decide which items deserve your own attention'. It explains why this is useful (confident answers can be trusted) and how it maps back (returns index). It does not list exclusions or alternative tools by name, but the usage niche is clear. Missing explicit 'when not to use' guidance prevents a 5.

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