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

Annotations already declare readOnlyHint, idempotentHint, and openWorldHint=false. Beyond that, the description adds genuinely useful context: it discloses that only 'non-secret' config is reported (a security/privacy guarantee) and that it 'does not make a model request' (a side-effect boundary). These go beyond the annotations, though it stops short of explaining permissions or error modes when credentials are missing.

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, tightly front-loaded with the core action, followed by the crucial side-effect clarification. No filler, every sentence earns its place.

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-param read-only diagnostic with annotation coverage, the description supplies what matters: what is reported, what is verified, what is not done. Without an output schema it does not describe the response shape (e.g., per-provider pass/fail fields), which is the only remaining gap for an agent calling it.

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?

Zero-parameter tool, which is the baseline-4 case. There are no parameters to misinterpret, and the empty schema is fully consistent with a no-arg health check.

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 specific operations: report non-secret configuration and verify configured provider credentials. The verb+resource is clear. It does not explicitly name or distinguish itself from the session-oriented siblings (get/consult/update/reset), relying on the reader to infer it is a distinct diagnostics tool rather than an advisor-session interaction.

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 says what it does but gives no guidance on when to prefer it over siblings or when a health check is warranted. There are no exclusions or alternatives named, leaving the agent to infer the use case from the tool name alone.

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?

Both sentences of the second half add real context beyond the annotations: 'the advisor cannot inspect your workspace' explains why the evidence pack must be self-contained, and 'Each call counts toward the per-session call budget' discloses a cost constraint that annotations do not convey. It stops short of describing what form the advice comes back in or what the budget actually is, and annotations already cover the read-only/destructive profile.

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, front-loaded with the decision triggers, then the input constraint, then the cost constraint. Every sentence earns its place with no restatement of the name or title.

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 15-parameter, open-world consultation tool with no output schema, the description covers the essential decision context: when to call, that the advisor is blind to the workspace, and that calls consume budget. It leaves unstated what the response looks like and what the budget limit is, minor gaps against otherwise rich schema documentation.

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 73%, so most fields (goal, state, attempt, evidence, context, question, options) are already documented in the schema. The description contributes only the high-level framing of 'a compact evidence pack'; it says nothing about kind selection, session_id reuse, include_history, or the undocumented fields such as constraints, style, files_touched, and max_history_messages. Baseline 3 applies when the schema does the heavy lifting.

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 ('Get an independent second opinion') backed by an advisor entity, which is clearly distinct from session-management siblings like get_advisor_session or reset_advisor_session. It never names those siblings explicitly, so the differentiation is implicit rather than spelled out.

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 three concrete triggering conditions: before committing to a non-obvious plan, after repeated failures, and before declaring complex work complete. That is strong positive guidance, but it does not state when NOT to consult or which sibling to use for session inspection versus a fresh consult.

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=false, and closed-world scope, so safety is covered. The description contributes the shape of what is returned (goal, constraints, decisions, open questions, call count, recent consults), but says nothing about how many "recent consults" are included, behavior for an unknown/expired session_id, or whether state persists across resets.

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?

One sentence, front-loaded with the verb, with the returned fields enumerated as a compact list. No filler or repetition of the title.

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 correctly takes on the job of enumerating return values, which it does thoroughly. The remaining gap is operational: how to obtain a valid session_id and what an unknown id yields.

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 single session_id parameter has 0% schema description coverage, so the description carries the burden of explaining it — yet it only restates "for a session_id" without saying where the id comes from (returned by consult_advisor? update_advisor_session?) or its format. It adds no meaning beyond the schema's type and length bounds.

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 ("Return") and resource (accumulated session state for a session_id) and enumerates the exact fields returned, so an agent can tell it apart from consult_advisor, update_advisor_session, and reset_advisor_session. It stops short of explicitly naming those siblings, but the read-only state-recall framing is unmistakable.

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 explicit when-to-use, when-not-to-use, or alternative named. Usage is only implied by the field list (you call it to recover prior context before consulting), leaving the agent to infer that this is the recall step in the advisor workflow.

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 destructiveHint=true, idempotentHint=false and readOnlyHint=false, so the safety profile is covered; the description usefully adds what exactly is lost (stored state and call counter) rather than leaving it to the agent to guess. It omits whether the deletion is recoverable and what happens to the session afterwards, but the core destruction scope is disclosed.

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 tight sentences with the destructive action front-loaded and the usage condition second. Every clause earns its place and nothing is padded.

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 destructive, non-idempotent mutation with no output schema, the definition tells the agent what is removed and when to invoke it, but not the outcome for an invalid or already-reset session_id or whether the advisor session can be re-established. Adequate, with visible 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?

Schema description coverage is 0% for the single required parameter, so the description must carry the burden, and it only restates the parameter name ('for a session_id') without explaining where the value comes from, its expected format, or whether an unknown id errors or is a no-op.

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 ('Delete the stored state and call counter') and scopes it to a single session_id. The delete semantics inherently separate it from get_advisor_session, update_advisor_session and consult_advisor, 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?

'Use when the task is complete or the context has changed fundamentally' gives a concrete trigger condition that an agent can act on. It stops short of stating when not to use it (e.g. mid-conversation, or before harvesting results) or what to do instead of resetting.

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 idempotent=true, destructive=false, openWorld=false, so the safety profile is covered; the description usefully adds that writes are 'durable' and cost-free relative to an advisor call. It omits what happens with an unknown session_id, whether omitted fields are preserved, and any confirmation of returned state.

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?

A single sentence with no filler, and the value proposition ('without spending an advisor call') is placed after the action so the agent reads what it does first.

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 5-parameter mutation tool with 0% schema coverage and no output schema, the description covers the field vocabulary but leaves merge semantics and failure behavior unexplained. Adequate to attempt a call, but not enough to call it confidently in edge cases.

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 0%, so the description must carry parameter meaning, and it does map loosely onto four of the five parameters (goal, constraints, touched files, open questions). However, it never clarifies append-vs-replace behavior (add_constraints vs replace_open_questions) or session_id's role, so the semantics only come from parameter names.

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?

Uses a specific verb ('Record') with the concrete state fields it writes (goal, constraints, touched files, open questions), and the clause 'without spending an advisor call' implicitly separates it from consult_advisor. It never says 'session' outright and doesn't name reset_advisor_session or get_advisor_session as siblings, so differentiation is inferred rather than stated.

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?

'Without spending an advisor call' gives a clear selection rule: use this for cheap state persistence instead of invoking the advisor, which points an agent toward consult_advisor as the alternative. It stops short of explicit when-not guidance, such as not using it to retrieve or reset state.

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