LLM Configurator

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

Annotations already establish read-only, idempotent, non-destructive, closed-world behavior, so the safety profile is covered. The description adds genuinely useful context beyond the annotations: it describes the return shape (fit verdict, memory arithmetic, decode-speed range with a confidence label) and the underlying model (memory sized from all parameters, speed from active parameters). There is no mention of caching, rate limits, or how the confidence label is derived, so it is not exhaustive, but it is well above the annotation baseline.

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 tight sentences with no filler: usage trigger first, then required inputs, then return shape and computational basis, then exclusions. Every sentence earns its place and the most important routing information 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?

The tool has an output schema, so the description is not obliged to explain return values, yet it still sketches what comes back (verdict, memory arithmetic, speed range with confidence label). Combined with annotations, required-parameter documentation, and sibling differentiation, an agent has everything it needs. Minor gap: it does not explain how quantisation and context_length affect the answer, though the schema gives defaults and ranges.

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, their defaults, and constraints. The description only names the two required ids that come from search_catalog and does not add syntax or format detail for quantisation or context_length. With the schema carrying the full parameter burden, this is the 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?

The description states a specific verb+resource ('check whether one specific model runs on one specific machine') and explicitly scopes it to a single hardware/model pair, distinguishing it from siblings like models_for_hardware (many models on one machine) and compare_hardware (side-by-side). It also names the required inputs and the computational basis (memory from all parameters, speed from active parameters), so the agent knows exactly what this tool 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?

The opening sentence gives an explicit trigger ('Use this when the user asks whether one specific model runs on one specific machine, at what memory cost, and roughly how fast'), and the closing sentence states exclusions: 'It does not recommend purchases and does not run anything.' Both when-to-use and when-not-to-use are covered without requiring inference.

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, and destructiveHint=false, so the safety profile is covered. The description adds genuine behavioral context beyond the annotations: the exact fields returned (memory, bandwidth, memory type, release year, status, dated price), the meaning of a null price, and a negative scope boundary. It doesn't cover cardinality edge cases or ordering of results, but it is well above the annotation baseline.

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 tight sentences with no padding; the usage trigger and prerequisite are front-loaded, followed by return fields and the scope exclusion. Slight redundancy between the first sentence's 'two to four machines' and the schema, but it earns its place as a routing cue.

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?

An output schema exists, so return-value explanation is optional, yet the description still names the compared fields and the null-price convention, which is a useful addition. Combined with annotations covering safety and the schema covering parameter constraints, the definition is complete for correct invocation; only minor details like result ordering are absent.

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% and the schema itself documents the 2-4 cardinality and the pattern via constraints. The description repeats the 'two to four hardware ids from search_catalog' framing already in the schema description, so it adds no syntax or format meaning beyond what structured fields provide. Baseline 3 is correct.

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 (compare) and resource (machines), specifies the 2-4 cardinality, and explicitly distinguishes itself from the sibling check_hardware_fit by stating 'It does not check any model against the machines'. An agent can route between the two tools without opening either 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?

It gives a clear when-to-use trigger ('when the user compares two to four machines'), a prerequisite (hardware ids from search_catalog), and names the alternative tool (check_hardware_fit) with the condition that selects it. All the routing information an agent needs is present.

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, openWorldHint=false, destructiveHint=false, so safety is covered. The description still adds value by disclosing the required input provenance (model_id must come from search_catalog) and the scope limit (does not answer hardware-fit questions). It does not mention rate limits or fallback behaviour, keeping it 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 tight sentences with no filler: trigger first, prerequisite second, return contents third, scope exclusion last. Each sentence carries distinct decision-relevant 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?

With an output schema present, annotations covering the safety profile, and a single fully documented parameter, the description supplies everything else an agent needs: when to call it, what id to supply, what it returns, and which sibling to use instead for hardware-fit questions.

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% on the single required parameter and the schema already states 'Model id from search_catalog.', which the description only repeats. Per the high-coverage baseline, 3 is correct since the schema carries the parameter semantics.

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 (get a model's specifications) and enumerates the exact attribute domains it covers: size, architecture, context window, licence. It also explicitly carves out the sibling boundary against check_hardware_fit, so an agent can route without opening either 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?

Gives an explicit trigger ('when the user asks about one model's size, architecture, context window or licence'), a prerequisite (a model id from search_catalog), and a clear exclusion with the named alternative ('It does not say whether the model fits any machine; use check_hardware_fit for that').

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 read-only, idempotent and closed-world, and the description adds real substance on top: the result is capped at 20 ranked models with a fit verdict and speed range, only one quantisation and context length apply per call, and non-fitting models are excluded. That scope disclosure is exactly what an agent needs and is not derivable from the schema.

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, all load-bearing, with the trigger and prerequisite front-loaded before the return shape and the coverage caveats. 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?

With a full input schema, an output schema and safety annotations already present, the description supplies the remaining context: the trigger condition, the prerequisite, and the ranking/fit/speed semantics of the result. 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 every parameter including defaults and enum values is already documented. The description adds only a light gloss ("optionally filtered to one use such as coding") and mentions the single-quantisation/context constraint, which the schema already conveys through its enum and defaults.

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 in user-intent terms: list the models a given machine can run. The direction is clearly the inverse of check_hardware_fit, and it names search_catalog as the source of the hardware id, but it never explicitly contrasts itself with those 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 an explicit trigger ("Use this when the user asks what models a given machine can run") plus a hard prerequisite (hardware id from search_catalog). It does not say when to prefer check_hardware_fit or compare_hardware instead, so it stops short of full alternative routing.

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 read-only, idempotent and closed-world, so the safety profile is covered. The description adds real behavioral context beyond them: the corpus is a fixed catalogue (closed world confirmed, not web), matching includes common aliases, and the result deliberately omits specs/fit/speed. Nothing about pagination or result shape, but the output schema covers returns.

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, no filler. The usage trigger and prerequisite come first, followed by the negative scope, so the most decision-relevant information 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?

An output schema exists, so return values need no explanation; the description instead covers the two things the schema cannot: when to reach for this tool first and what it deliberately will not return. Nothing needed 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 baseline is 3, and the schema already supplies enum values, bounds and example queries. The description earns above baseline by disclosing matching semantics the schema cannot express: names and common aliases are matched against a fixed catalogue, which tells the agent how to phrase the query.

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 (search/find) and resource (a fixed catalogue of local-LLM hardware and models), and explicitly names the trigger condition: the user names a GPU, Mac, mini PC or language model and you need its id. It is clearly distinguishable from get_model_specs, check_hardware_fit and compare_hardware.

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?

Front-loads 'Use this first when...' and states the prerequisite (you need an id for the other tools), which routes the agent ahead of the spec/fit/compare siblings. It also gives explicit exclusions: no specs, no fit or speed, no web search.

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