ConferenceCaptioning

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

Annotations already declare read-only, idempotent, non-destructive, open-world, so the description adds the non-obvious: translations are a re-translated window rather than a stored record, cursor applies only to spoken-language text, output is wrapped in <untrusted-text> as data, and the text is forwarded to the calling model.

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, front-loaded with what is returned, then cursor behavior, then the safety/prompt-injection caveat. Dense but each sentence carries distinct information; the final sentence is slightly abrupt.

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 return-value burden and does so: it describes the text content, the cursor to echo back, the <untrusted-text> wrapping, and the fact that output goes to the model. Nothing needed to invoke it safely 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 coverage is 100%, so the baseline is 3; the description still adds meaning by clarifying that the cursor only advances spoken-language text and that translation languages behave differently from the source stream.

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 ('newest live caption text of a stream') plus scope (spoken language or a translation language), which cleanly separates it from get_saved_transcript and get_stream_context among 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 a concrete usage rule: pass the returned cursor back to receive only new spoken-language text, and notes the 'source' default for language. It never explicitly states when to prefer get_saved_transcript over this tool, so routing guidance is clear but not complete.

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 readOnly, idempotent, and non-destructive behavior, so the safety profile is covered. The description adds genuine context beyond that: the default page size of 20,000 characters, that total length and next offset are returned, and that text is wrapped in <untrusted-text> and forwarded to the model — an important trust-boundary disclosure. It stops short of describing behavior for edge cases like missing transcripts.

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?

Front-loaded with the core purpose and page size, followed by return-value and trust caveats. Sentences are short and mostly earn their place, though the final 'Returned text is sent to the AI model you are using' is somewhat redundant with the preceding untrusted-text warning.

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 responsibly explains the return shape (total length, next offset, wrapped text). But for a 5-parameter tool with 40% schema coverage, key semantics — how to choose a stream, what includeHidden reveals — are absent, so an agent could call it correctly only for the common single-stream case.

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 only 40% schema description coverage, the description partially compensates by explaining the default page size (limit) and the returned next offset (offset), plus the <untrusted-text> wrapper. But it says nothing about streamName or includeHidden, which are undocumented in the schema as well, leaving half the parameters ambiguous.

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: retrieves the text of one saved transcript, in pages. It is distinguishable from list_saved_transcripts (which enumerates transcripts) by its singular focus and its mention of paginated text. It does not explicitly name or contrast with sibling tools like get_live_captions, so it falls 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?

The mention that 'the result says the total length and the next offset' implies the pagination workflow, so an agent can infer to call repeatedly with offset. However, there is no explicit guidance on when to use this versus get_live_captions or how to obtain the id (that only appears in the schema's id description). Usage is implied rather than stated.

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 readOnly, idempotent, non-destructive and openWorld, so the bar is lower. The description adds two things beyond that: the data is public (no auth concern) and, notably, that the returned text is sent to the AI model in use — a genuine data-flow disclosure an agent should know. It stops short of describing the return shape or caching behavior.

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, resource identity front-loaded, with the privacy note placed last where it belongs. Terse but not wasted; the fragment style costs it a point only slightly.

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, non-destructive read with rich annotations this covers purpose and privacy, but it never indicates what the response looks like (field-level design data vs glossary term structure) and gives no selection guidance against the four sibling read tools.

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 single optional streamName parameter already documents its default fallbacks (CONFCAP_STREAM, owner's only stream). The description adds no parameter-level detail, so this sits at the schema-covered baseline.

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 the concrete resources returned — a stream's design (colour, logo) plus the glossary terms it uses — which implicitly separates it from get_stream_status and list_my_streams. It lacks an explicit verb ("get"/"return") and never names a sibling, so differentiation is left 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?

There is no statement of when to call this versus get_stream_status or list_my_streams, and no prerequisites or exclusions. The agent must infer that design/glossary lookup is the trigger.

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, openWorldHint, and destructiveHint=false. The description adds the possible return states and notes that returned text is sent to the AI model, but it does not clarify what 'returned text' refers to, nor does it explain auth flow 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?

The description is two sentences and front-loads the core question of whether a stream is live and readable. The model-exposure note is relevant, though the phrasing 'Returned text is sent to the AI model' is slightly ambiguous.

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?

This is a simple read-only status tool with a fully described single parameter and no output schema. The description lists possible state values but does not describe the full return shape, such as language or readability fields, so it is adequate but leaves gaps.

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?

There is one optional streamName parameter, and schema description coverage is 100%, including its default behavior. The description adds no additional parameter semantics, so the baseline of 3 is appropriate.

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 purpose: checking whether a stream is live, what languages it uses, and whether it can be read, including the possible status values. This clearly distinguishes it from siblings like get_live_captions, get_saved_transcript, or get_stream_context, though it does not explicitly name those alternatives.

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 implies usage for checking stream availability/readability before other operations, but it does not explicitly say when to use this tool versus alternatives. No exclusions or sibling-tool routing guidance 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?

Annotations already declare read-only, idempotent, non-destructive, and open-world behavior. The description adds useful auth scoping and a data-flow note that returned text is sent to the AI model, but it does not cover return format, pagination, 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?

The description is short and front-loads the core purpose. The second sentence about returned text going to the AI model is relevant but slightly tangential to the primary listing action.

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-parameter list tool with no output schema, the description covers the core action and auth scope but omits what the returned list looks like and whether pagination or filtering exists. More detail on the response shape would help an agent use it confidently.

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 has zero parameters, so the baseline is 4. There are no parameter semantics to clarify, and the description appropriately does not invent any.

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: 'List the streams' scoped to those the configured developer API key can read. It is clear what the tool does, but it does not explicitly distinguish itself from sibling tools such as get_stream_context or list_saved_transcripts.

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 'streams that the configured developer API key can read,' suggesting the tool is for discovering accessible streams. However, there is no explicit guidance on when to use this versus related tools like get_stream_status or get_stream_context.

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?

The description adds substantial behavioral context beyond annotations: it discloses the need for a developer API key, specifies oldest-first ordering, warns that question text is untrusted and should be treated as data, and notes that returned text is sent to the AI model. These details are critical for safe and correct invocation.

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 front-loaded with the core purpose and ordering, then efficiently adds authentication and safety context. Every sentence earns its place without unnecessary repetition.

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 read-only list tool with fully documented parameters and annotations covering safety, the description supplies essential context including auth requirements, ordering, and untrusted-text handling. It does not describe the response shape beyond noting the returned text is sent to the AI model, but no output schema exists and the omission is minor.

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 the two optional parameters thoroughly. The description adds no parameter-specific semantics beyond what the schema provides, which is appropriate for a baseline score.

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 that this tool retrieves audience questions from the stream page, including ordering as oldest first. It does not explicitly differentiate the tool from siblings such as get_live_captions or get_stream_context, so it falls 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?

Usage is implied by the resource description, and the need for a developer API key is noted. However, there is no explicit guidance on when to use this tool instead of related siblings, nor are there any stated exclusions.

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 the safety profile (readOnly, idempotent, non-destructive), and the description adds genuinely non-obvious behavior: developer API key requirement, partial history coverage, hidden transcripts gated behind includeHidden, spoken-language-only content, and a data-handling note that returned text is sent to the calling model. That last point is unusual and useful, though nothing is said about result volume or ordering.

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?

Front-loaded with purpose and the field list, then a short chain of one-clause caveats — no filler. The telegraphic style (sentence fragments) is efficient but slightly reduces readability for an agent parsing constraints.

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 two-parameter read-only list tool with no output schema, the description covers purpose, return shape, auth, availability, and the hidden-transcript gate — close to fully self-sufficient. Only ordering/pagination expectations are unaddressed, a minor gap for a list operation.

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?

streamName's defaulting rules are already in the schema, but includeHidden has no schema description at all, and the description compensates by explaining that hidden transcripts appear only when it is set. That closes the 50% coverage gap rather than restating structured data.

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?

Opens with a specific resource and scope: 'The owner's saved transcripts for a stream', and enumerates the returned fields (id, title, note, times, length). That is enough to distinguish it from the singular sibling get_saved_transcript and from list_my_streams, though neither alternative is named explicitly.

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 preconditions (needs a developer API key) and availability caveats ('Only some streams store history', spoken language only), which shape when the call is worthwhile. However, it never states when to use this versus get_saved_transcript or how it relates to list_my_streams, so the routing decision is only implied by the plural name.

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