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

Goes well beyond the annotations: it discloses the intermediate cancel_pending state pending ops confirmation, the refund behavior for a confirm before dialing, the specific cancel_too_late error when the dial already started, and the billing/VA-payment consequence of a dial after a confirmed cancel. None of this is derivable from destructiveHint/idempotentHint/openWorldHint.

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?

Five short sentences, front-loaded with the action, then packed with state, refund, error, and billing facts. No filler sentences and no repetition of the title or annotations.

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 not be described, and the description covers lifecycle states, the failure mode, and financial consequences thoroughly. Minor gap: no mention of permissions/ownership or who may cancel, which for a destructive operation would be useful context.

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 required parameter (call_id) with 0% schema description coverage, and the description never mentions it. The parameter is largely self-evident from the name, but the description adds no meaning (format, whether it is a UUID, whether the call must be in a particular state).

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 ('Cancel a call') and immediately elaborates the state the call lands in, so the agent knows exactly what operation this is. It does not explicitly contrast itself with the call-vs-task siblings (cancel_task, confirm_task), but the resource is unambiguous from the name and first sentence.

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 rather than routed: the description explains what happens depending on call state (pre-dial refund, dial-started failure), which helps the agent predict outcomes, but it never names an alternative tool or states prerequisite conditions for choosing this over cancel_task or update_call.

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 and idempotentHint=true, but the description goes well beyond them by spelling out the concrete effects: the hold is released, billing is limited to tracked time, and no completion webhook fires. That is exactly the 'what gets destroyed' context an agent needs before a destructive call.

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?

Five short, front-loaded sentences with no filler; each effect (hold release, billing, webhook suppression) earns its place. The ambiguous 'Customer never mind' opener costs a point because it reads as noise rather than usable instruction.

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 rich annotations, a full output schema, and 100% parameter coverage, the description only needs to add behavioral context, which it does for effects and preconditions. The unexplained 'Customer never mind' phrase is the one loose end that keeps it from fully closing the loop.

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 all three parameters (reason, taskId, idempotencyKey) are already documented in the schema. The description only restates that Idempotency-Key is required, adding no syntax or format detail beyond what the schema provides, so the baseline 3 applies.

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 (cancel a task) and adds concrete business semantics: releases the hold, bills only tracked time, suppresses the completion webhook. The opening fragment 'Customer never mind' is an odd preamble but the remainder clearly distinguishes this from confirm_task/reopen_task, which move a task forward rather than ending it.

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 precondition ('Ask the human first'), which is real when-to-use guidance for a destructive operation. It does not name alternatives such as confirm_task or reopen_task, so an agent must still infer the surrounding workflow, keeping it short of a 5.

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 and idempotentHint, but the description adds substantial context beyond them: the precise state transition, that tracked hours are settled and the board and ledger hold released, the 3-day auto-confirmation fallback, and the human-approval expectation. The idempotency note explains the mechanism ('a repeat of the same key returns the same result') rather than merely restating the hint.

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 the tool does and the state it applies to, then effects, then the human-approval caution and idempotency contract. No filler or redundant restatement 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?

An output schema exists so return values need no explanation, and annotations cover the safety profile. The description still supplies everything an agent needs to decide and act: precondition, side effects, human-approval expectation, timing fallback, and the required idempotency key contract.

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 parameters are already documented, and the baseline would be 3. The description adds genuine value on idempotencyKey by stating it is required and describing its repeat-call semantics, though taskId gets no description-level elaboration beyond the schema.

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 ('Accept a delivery') plus the exact precondition state ('awaitingCustomerConfirm'), and enumerates the resulting effects (settles hours, releases board and ledger hold, marks completed). This distinguishes it from siblings like cancel_task and reopen_task, which operate on different lifecycle states, without needing to name them.

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 clear operating context: only applies to tasks in awaitingCustomerConfirm, and instructs 'Ask the human first.' It also explains the fallback ('the job confirms itself after 3 days'), which tells the agent this is not always strictly necessary. It stops short of explicitly contrasting with sibling alternatives such as cancel_task or reopen_task, so it's clear context rather than full when/when-not 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?

Goes well beyond the annotations (readOnlyHint=false, destructiveHint=true, idempotentHint=true, openWorldHint=true) by disclosing the asynchronous lifecycle: the job is not started until the customer clicks Start, and the receipt status is awaiting_customer_start, plus a taskId is returned. It also adds a spend/confirmation policy and a secrets-handling prohibition that the annotations cannot express.

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?

Dense but nearly waste-free, with each sentence carrying a distinct rule (prerequisite, purpose, use cases, safety, return, lifecycle). It loses a point for leading with the get_balance/confirmation precondition before stating what the tool actually does, which weakens front-loading of the core purpose.

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 complex, destructive, open-world creation tool with an output schema, the description supplies everything needed to invoke it correctly: prerequisites, confirmation policy, scope of use, secret-handling limits, the returned taskId, and the fact that work does not begin until customer activation. Nothing material 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 88% and 8 parameters are present, so the schema already documents fields such as maxHours, idempotencyKey, and description in detail. The description only indirectly touches the spend-cap parameter via the get_balance/confirm-for-real-hours guidance and does not explain any parameter format beyond what the schema provides, so the baseline 3 applies.

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 (delegate) and resource (task to a real human VA) and adds the governing condition (when the AI cannot complete it autonomously). This cleanly separates it from siblings like list_tasks, get_task, cancel_task, and place_call, which are retrieval/control operations rather than task creation.

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 explicit when-to-use categories (phone calls, owner-authorized login/captcha, physical-world actions, subjective QA, manual data entry) and explicit exclusions (never send passwords or OTP). It also names the prerequisite sequence: call get_balance first, then confirm with the user before sending, unless already authorized.

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 read-only, idempotent and non-destructive, and the description adds genuinely new behavioral facts: addFundsUrl is POST /api/v1/billing/add-funds yet this read creates no Stripe session, and canDelegate semantics for the zero-hours / no-room cases. It does not cover auth requirements or rate limits, so not 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?

Front-loaded with the resource and its returns, then the canDelegate rules and the invocation cue. Efficient overall, though the field-by-field wallet breakdown and the addFundsUrl/Stripe detail are somewhat dense and partly overlap with the output 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?

The schema has no params, annotations carry the safety profile, and an output schema exists so return values needn't be re-specified. The description still supplies the non-obvious semantics an agent needs: how canDelegate false should be interpreted, and that this read has no billing side effect.

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 input parameters, so there is nothing parameter-side to compensate for; baseline is 4. The description instead spends its detail on return-field semantics (balance_cents vs wallet_cents vs holds_cents), which sits outside this dimension.

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, naming exactly what comes back: remaining VA hours, capRoom, whether a new cap can be filed, and the calls wallet breakdown. It also disambiguates itself from siblings by asserting 'Does not create a task' and 'not a buy-hours signal', so an agent can separate it from place_call/delegate_task without opening a 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 clear invocation context: 'Call this before proposing spend.' It also sets two boundaries — not a buy-hours signal, and it does not create a task. It stops short of naming an alternative tool for the buy-hours or cap-filing paths, so it is strong context rather than full when/when-not 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 readOnlyHint=true, idempotentHint=true, destructiveHint=false, and openWorldHint=true. The description adds useful context that the response may include needs_info missing_fields or the final result, but it does not discuss permissions, rate limits, or other operational 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?

The description is a single front-loaded sentence with no wasted wording. It immediately states the action and resource, then adds the most relevant result detail.

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 is simple, annotations are rich, and an output schema exists, so the description does not need to explain return values. Still, it leaves gaps around call_id sourcing and does not distinguish this tool from get_transcript, which an agent may need to choose correctly.

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 required parameter, call_id, but the schema description coverage is 0% and the description does not explain where call_id comes from, its format, or any constraints. The description adds essentially no meaning beyond the schema's parameter name.

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: 'Read one human phone call.' It also names the kind of content available, such as needs_info missing_fields or the final result. However, it does not distinguish this from sibling tools like get_transcript or update_call.

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 use for retrieving a single call by ID, but gives no explicit when-to-use guidance, prerequisites, or alternatives. It does not say when to choose this over get_transcript or other call-related siblings.

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 establish readOnly/idempotent/non-destructive, but the description adds non-obvious behavior: a customer cancel lands as 'cancelled' rather than pending/in_progress and does not fire the completion webhook. Those are side effects an agent could not infer from the annotations or schema. It stops short of covering rate limits or response shape beyond the listed fields.

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-loads the returned data and then layers the interpretation rules (awaitingCustomerConfirm, amNote, deliverableLinks, terminal statuses, cancel semantics) in short, terse sentences. A couple of the field glosses ('amNote is the AM or VA note') are near-redundant given the output schema, but overall every sentence carries usable 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?

For a single-parameter read tool with an output schema and full annotation coverage, this description supplies everything an agent needs: what is returned, how to read key fields, what to do on awaitingCustomerConfirm, when to stop, and the webhook quirk for customer cancels. Nothing material 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?

Only one parameter, and the schema already documents it at 100% coverage ('The taskId returned by delegate_task'). The description adds no extra syntax, format, or sourcing guidance for taskId beyond what the schema says, so the baseline 3 applies.

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 verb+resource ('Get the current status, title, completion timestamp, and full result notes for a specific delegated task') and enumerates the returned fields, so it is immediately distinguishable from list_tasks or get_call. The scope is a single task identified by taskId, and that is stated plainly.

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 explicit routing: when awaitingCustomerConfirm is true, ask the human and then call confirm_task, cancel_task, or reopen_task, naming the sibling tools by name. It also states the stopping condition (status completed, failed, or cancelled), so the agent knows when not to keep polling.

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/openWorld, and the description goes beyond them by enumerating the distinct outcome states (transcript_not_ready, 'no transcript: call not placed', call_cancelled) and the source/recorded fields. It also pre-empts a likely false expectation by stating a recorded transcript is not returned.

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-loads the core purpose and the negative ('not a recording') in the first two sentences, then uses short sentences for each state. Dense with information and mostly waste-free, though the trailing 'may be a separate paid feature later' is speculative roadmap detail rather than call guidance.

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 usefully explains what the result contains and its error/edge states. The remaining gap is the call_id parameter's semantics, which nothing else in the structured data supplies.

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% and the single required call_id has no description, so the description must compensate. 'For one call' implies the parameter is a call identifier but adds no format, source, or where-to-find-it guidance, leaving the agent to guess acceptable values.

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 (read) and resource (VA wrap-up summary for one call) and explicitly negates the confusable interpretation: 'This is not a recording.' That distinguishes it from a recording/audio retrieval, though it does not explicitly contrast with sibling get_call, which also takes a call_id.

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 clear conditional context: usable once the call is wrap_up or completed, otherwise transcript_not_ready. It does not, however, route the agent between this and sibling get_call/update_call, leaving the tool-selection question partly to 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 read-only, idempotent, non-destructive behavior, so the description correctly spends its words elsewhere: it explains the awaitingCustomerConfirm gate (delivery done, human must confirm/cancel/reopen) and the terminal nature of 'cancelled'. It does not mention any bound on 'recently' (result cap or pagination), which is the one behavioral trait still 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?

Four short sentences with the operation and returned fields front-loaded, followed by the interpretive notes an agent needs to read the flags correctly. No filler, though the field-by-field gloss overlaps with the output schema and could be tightened.

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 and annotations covering the safety profile, the description only needs to add interpretation the structured data cannot: what awaitingCustomerConfirm means, how status buckets exclude cancelled jobs, and that 'cancelled' is terminal. For a one-parameter list tool this is sufficient, with pagination bounds the only minor omission.

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% across the single status parameter, so the description carries the burden and partially meets it by defining the semantics of 'pending', 'in_progress' (both omit customer-cancelled jobs) and 'cancelled' (terminal). 'completed', 'failed', and the 'all' default remain unaddressed, so the enum is not fully covered.

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 ('List recently delegated tasks for this account') plus the shape of each returned row, which cleanly separates it from single-task siblings like get_task or confirm_task. An agent can identify both the operation and the scope without consulting the 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?

Usage is implied by 'recently delegated tasks for this account', but the description never states when to prefer this over get_task or the mutation siblings (confirm_task, cancel_task, reopen_task), nor does it mention any prerequisite. Adequate for a simple listing tool, but no explicit routing guidance.

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?

Goes well beyond the annotations by disclosing the billing/hold model (first-minute tier minimum, per-minute thereafter, $6.50 hold at the 5-minute default), the interaction between max_minutes and the hold, the hard rejection of record_call=true, and the live-market restriction. These are exactly the operational facts (cost exposure, irreversible spend, geographic limits) an agent needs before a destructive, open-world write.

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 action, then the cost model, then constraints and alternatives. Nearly every sentence carries a distinct fact, though the terse fragment style ("Only US and CA are live.") makes the block dense to parse at 24 params.

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 and annotations carry the safety profile, so the description needn't cover returns. It covers the highest-risk constraints (spend, idempotency, recording rejection, market and business-hour rules), leaving only peripheral fields undocumented.

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 24 params at 58% schema coverage, the description compensates by specifying the questions object shape, its 80/1000-char and 20-item limits, the 4000-char goal cap, required idempotency key, and the pricing/semantics of max_minutes. It leaves several lower-risk params (script, purpose, language, script_mode, shareable_contact, etc.) to the schema.

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 ("Place one human phone call") plus scope limits (US and CA only, one call, human VA). It also explicitly contrasts itself with sibling validate_call, so an agent can distinguish the two 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?

Names the polling path (get_call) and callback_url alternative, and clarifies that validate_call checks the body without placing a hold, which is clear routing guidance. It stops short of an explicit when-not/preconditions statement (e.g. that consent_attestation must be true before placing).

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 as a non-readonly, idempotent, non-destructive, open-world mutation. The description adds real behavioral context beyond them: 'Does not settle hours' rules out a side effect an agent might otherwise assume, the required Idempotency-Key is called out, and human confirmation is mandated before 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?

Four short sentences, each carrying a distinct fact: the state transition, the destination, the side-effect exclusion, the human gate, and the idempotency requirement. The precondition and action are front-loaded with no 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?

An output schema exists, so return values need no explanation, and the safety profile is covered by annotations. For a 3-parameter mutation the description covers precondition, side effects, human approval, and idempotency; it stops short only of describing failure behavior when the task is not in awaitingCustomerConfirm.

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 all three parameters (taskId, reason, idempotencyKey) are already documented in the schema, including the required markers and 'do not include secrets' on reason. The description only echoes the Idempotency-Key requirement, adding no new parameter meaning, so the baseline 3 applies.

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 (send back / reopen), the resource (a delivery), and the precise precondition state (awaitingCustomerConfirm) plus the destination (account manager). An agent can distinguish this from confirm_task or cancel_task without opening either schema, because the state and direction of the transition are named.

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?

'Ask the human first' gives an explicit gating condition, and the awaitingCustomerConfirm state narrows when the tool applies. It does not explicitly name the sibling alternatives (confirm_task, cancel_task) that an agent might otherwise pick, so it falls short of a full when-not/alternative statement.

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=false, idempotentHint=true, destructiveHint=false, and openWorldHint=true. The description usefully clarifies that this only fills MISSING fields (partial, non-overwriting) and that results mirror place_call, which is consistent with the idempotent, non-destructive profile. It adds little else about auth, side effects, or limits, so it earns a moderate score against an already-covered 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?

Two tightly worded sentences with the core action front-loaded and zero filler. It is efficient, though the extreme brevity is arguably under-specification given the 24-parameter surface.

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 24 properties at 8% coverage and a mutation tool, the description is far too thin on inputs; it does not explain which fields are actually fill-able or their formats. It does offload return-value explanation to the output schema and place_call, but the input side remains largely 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 only 8% across 24 parameters, so the description must carry the load and it does not: beyond 'missing_fields', it explains no parameter names, formats, or constraints. The result-shapes reference to place_call gives no input guidance, leaving the bulk of parameters undocumented in both schema and description.

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 verb and resource ('Fill missing_fields on a call') and adds a state precondition ('in needs_info'). It also ties output to the sibling place_call ('Same result shapes as place_call'), so the agent can tell it apart from place_call and cancel_call. It stops short of full sibling differentiation, so not 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 precondition 'on a call in needs_info' implies when to reach for this tool versus place_call (new call) or cancel_call, but it never explicitly states when to use it or when another tool is preferred. Guidance is inferable but not spelled out.

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 (readOnlyHint, idempotentHint, destructiveHint=false, openWorldHint), so the bar is lower. The description adds genuinely non-obvious behavior: the response carries hold_cents (with a concrete tier-1 default), whether recipient_timezone was inferred, and the counter-intuitive rule that valid stays true when spendable balance is below the hold, with an insufficient_balance warning instead.

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, front-loaded with the core distinction from place_call before moving to response semantics. Dense but every sentence carries information; the tier-1 pricing detail is slightly incidental but supports cost reasoning.

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 not be fully explained, yet the description helpfully covers the key conditional fields (hold_cents, valid, warnings.insufficient_balance). Given the high parameter count and near-zero schema coverage, the main remaining gap is parameter-level guidance, which the place_call cross-reference only partially bridges.

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 are 23 parameters at only 9% schema description coverage, so the description carries the burden of explaining them. It only notes the absence of idempotencyKey and cross-references place_call for the rest, leaving nearly all fields (purpose enum, time_window, consent_attestation, retry_preference, etc.) undocumented in either place.

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 ('Check a phone-call body') and immediately distinguishes itself from the sibling place_call by stating it does not create a call or place a hold. An agent can tell validate_call apart from place_call 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 dry-run framing ('without creating a call or placing a hold') gives clear context for when this tool is appropriate versus place_call. It stops short of an explicit instruction such as 'call this before place_call to preview cost and validity', so the routing is strongly implied rather than spelled out.

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