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 substantive context on top: what content comes back (answers or transcripts, criterion scores, feedback, verified practice/review links), pagination via continuation for longer resources, and explicit exclusions (raw audio, payment data, internal diagnostics) plus the guarantee that no new answer is scored. That exclusion list and the idempotency-adjacent note about reusing existing evidence are genuine value beyond the annotations; only rate limits or failure modes are absent.

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 return contract is front-loaded, which is good, but the prose is dense and contains low-yield phrasing such as "outside this resource contract" and the closing sentence about existing exact evidence providing "missing content, continuation, another version or an explicitly requested refresh," which is hard to parse on first read. The content is mostly relevant, yet several clauses could be tightened without losing meaning.

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 no output schema, the description carries the burden of describing return content, and it does so reasonably: resource types, included answer/transcript/score/feedback fields, embedded links, continuation, and explicit exclusions. For a two-parameter read-only fetch this is close to sufficient; the main remaining gap is the absence of any concrete notion of the returned object's shape or how continuation tokens surface.

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 both `uri` and `cursor` are already documented, which sets the baseline at 3. The description adds a real constraint the schema only hints at: the URI must be one returned by search from a specific family, and continuation is framed as the mechanism for longer resources. It does not add syntax or format detail for the cursor itself, so it earns a modest lift rather than a full 5.

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: it returns exactly one Joe Speaking question, saved speaking session or collection item keyed by a returned URI, and it enumerates the supported URI families (question-bank, live-conversation, recording-practice, collection-item). It separates retrieval from practice ("Speaking practice, recording, scoring and mock tests take place in Joe Speaking"), though it never names the sibling tools like `search` or `open_joe_speaking` directly, so differentiation is inferred rather than explicit.

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 clear when-to-use guidance: review saved fluency, vocabulary, grammar or pronunciation feedback, inspect a past answer or cue card, or continue reading a returned resource. It also gives a when-not signal by pushing practice, recording, scoring and mock tests to Joe Speaking, and clarifies that existing evidence in the conversation may suffice unless the agent needs missing content, continuation, another version, or an explicit refresh. It stops short of pointing at a named alternative tool, which keeps it from a 5.

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