Sternly

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

Annotations already declare readOnlyHint, idempotentHint, non-destructive and closed-world, so the safety profile is covered. The description adds real behavioral context on top: the response is staged (status first, result only after completion) and the result includes a signed evidence package link. It does not mention what happens for an unknown or malformed check id, which is the main remaining gap.

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, front-loaded with what the tool returns and followed by the one procedural instruction that matters. Every clause carries information; nothing is repeated from the title or schema.

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 takes on the job of describing return values and does so by listing status, headline, counts, depth, findings and the evidence link. For a single-parameter read tool that is nearly complete; only the volume or shape of findings and whether repeated calls are expected while running are left unstated.

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 has to carry the single parameter, and it does by tying check_id to the value returned by submit_claim_bundle — provenance the bare pattern ^chk_[0-9a-f]{32}$ cannot convey. It does not restate the format, which is fine since the pattern already covers it, but no error semantics for a stale or malformed id are given.

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?

Names a specific resource (a Sternly check) and states the exact information returned, enumerating the status vocabulary (running, done, failed) and the result fields (headline, counts, depth, findings, evidence link). It also distinguishes itself from submit_claim_bundle by naming it as the source of the id, so an agent can separate the read step from the submit step 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?

The description gives clear operational context: call this with the check id that submit_claim_bundle returned, and expect a result only once the status is done. That establishes the typical call sequence. It stops short of explicit exclusions or polling guidance (e.g. whether to re-call while running), so it is clear but not fully prescriptive.

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, closed-world, and non-destructive, so the safety profile is covered and the description's "do not change it" is largely redundant. What the description does add is the shape of the result (version, owner, effective date), which is genuinely useful context beyond the annotations, but it says nothing about behavior for terms that do not exist or whether omitting the argument returns the whole workspace.

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 short sentences with no filler, and the substantive content (the returned metadata) comes first. The opening fragment is slightly awkward but costs little.

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 and zero schema description coverage, the description carries most of the burden, yet it omits the behavior of the only parameter and the response for unknown or missing terms. It covers the tool's purpose and result fields but leaves the agent guessing at the actual call semantics.

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 parameter (terms, an array of strings capped at 50) with 0% schema description coverage, so the schema itself explains nothing about intended input. The description implies term lookup but does not say the parameter is a list of term names, nor what an empty or omitted value means.

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 conveys the content returned (workspace term definitions with version, owner, and effective date), but it opens with a noun phrase rather than a verb+resource statement of the action. It never distinguishes itself from the sibling propose_definitions, which is the closest-looking alternative and the one an agent is most likely to confuse it with.

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 guidance on when to call this versus get_check, propose_definitions, or submit_claim_bundle. The only usage-adjacent sentence, "Give the definition as it is; do not change it," is an instruction about handling the result, not a routing rule, and no prerequisites or exclusions are 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 declare this is a non-read-only, non-destructive, non-idempotent write, and the description adds meaningful context beyond that: proposals require human approval and nothing takes effect immediately. It still omits what happens on a duplicate term and what the proposer sees back, which matters given idempotentHint=false.

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 short sentences with the approval constraint front-loaded alongside the action. The second sentence largely restates the first's approval clause, so it is slightly redundant rather than maximally economical.

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 and the annotations cover only the safety profile, so the description should carry the full behavioral burden. It covers the staging/approval model well but leaves batch limits, duplicate handling, and the post-call state unexplained.

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% for a single required parameter that is an array of nested objects with term, definition, and note fields. The phrase 'in their words' usefully signals verbatim capture, but the description never explains the term/definition/note roles or the 20-item cap.

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 ('send a definition ... to the workspace') plus the outcome (human approval before use). It is clearly a write-to-staging operation, distinguishable from the read-oriented siblings get_definitions and get_check, though it never names those siblings 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?

The approval workflow implies when to use it (capturing a person's own definition for review), but there is no explicit when-to-use or when-not-to-use guidance and no reference to get_definitions or submit_claim_bundle as alternatives. The agent must infer the routing decision.

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?

Beyond the annotations, it discloses the return contract (a check id and a link to give the person), the async behavior with wait=false, and the integrity rules ('do not rewrite, shorten or add to it', 'do not mark which part came from where'). That is meaningful behavior an agent cannot read off readOnlyHint/destructiveHint. It does not mention idempotency_key, which matters given idempotentHint=false.

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 action, then integrity rules, then mode/producer, then return and wait semantics — a sensible ordering. A few clauses are dense ('do not mark which part came from where; Sternly reads the work itself'), but nothing is filler.

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 14-parameter tool with no output schema and 29% coverage, the description covers the required path and the return value but leaves several nested structures (layout, capture, tool_log, notes) unaddressed. An agent can make a valid minimal call, but the richer bundle fields require reading the schema.

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 29% schema coverage across 14 parameters, the description carries real weight: it maps text, request ('what the person asked for in their own words'), files, sources, programs with printed output, policy.mode and producer to their purpose. It also constrains producer ('name a producer only if you know who made it'). notes, layout, capture, tool_log and idempotency_key remain unexplained.

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: hand finished work to Sternly for an independent check. It also distinguishes itself from the sibling by saying 'get_check returns the result', so an agent knows this is the submission step, not the retrieval step.

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?

'after the fact' and 'finished work' bound when to call it, and get_check is named as the follow-up. It also tells the agent when to set wait=false. It stops short of explicit exclusions (e.g. not for work in progress), so it is clear context rather than full when/when-not guidance.

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