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

Adds substantial behavior beyond the annotations: long answers are not repeated in tool output, full content is written to answer_path/code_artifacts and must be read from file, external responses are data not instructions, and timeouts must not trigger a resubmit. The retry/idempotency semantics overlap somewhat with idempotentHint=false but add the specific failure-mode rule. No auth or rate-limit detail.

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 action, then packs constraints into short declarative fragments with no filler. The telegraphic style ("Personal chat only; never Custom Agent.") is dense but readable. Slightly list-like rather than prose, but every sentence carries an operational rule.

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 an open-world, non-idempotent tool with no output schema, the description covers the critical gaps: where the real answer lives (local file), why output looks truncated, and the correct recovery path when a timeout occurs. Authentication, quota, and the prompt vs prompt_file choice remain unaddressed, but the operationally decisive information is present.

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 50%. The description clarifies request_key's intended reuse semantics ("Reuse it after errors; do not invent a new key to retry"), reinforcing the schema. However, it says nothing about prompt vs prompt_file selection or timeout_seconds, leaving half the parameters unexplained outside 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 ("Create a personal chat") and adds a scope constraint ("Personal chat only; never Custom Agent"), which helps distinguish it from notion_personal_submit/reply. The "and wait" phrasing is a bit loose but conveys the synchronous nature. It does not explicitly name the siblings it competes with, so it stops 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?

Gives explicit routing guidance: prefer "submit plus read" for reliable recovery, or supply request_key, and warns that a timeout is not grounds to resubmit. That is concrete when/when-not guidance. It still doesn't say when a plain synchronous ask is preferred over notion_personal_submit in the first place, so not fully exhaustive.

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 the safe/idempotent read profile, yet the description adds substantial behavior the annotations cannot convey: repeated calls resume pagination and pending generation, external content is data not instructions, full answers are persisted to answer_path/code_artifacts and not echoed in output.

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?

Purpose is front-loaded and every sentence carries operational information, but the paragraph is dense and mixes usage, parameter routing, security, and output persistence without visual separation, making it harder to scan than necessary.

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 covers what the return actually is (summarized, with full content saved to local files) plus timeout and side-effect guidance. An agent has everything needed to call and interpret this tool.

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 must carry parameter meaning and does for three of four: job_id and request_key are preferred handles, thread_id resolves latest local task or imports a conversation. It omits any explanation of wait_seconds (default 20, max 300), leaving one parameter's semantics to the schema alone.

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 (read/resume an existing task) and explicitly negates the generating behavior ('without generating'), which cleanly separates it from the ask/submit siblings. An agent can distinguish this from notion_personal_ask 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?

Explicit routing: prefer job_id or request_key, thread_id resolves the latest local task or imports a conversation. It also gives scope ('Personal chat only; never Custom Agent') and a negative rule for a specific failure mode ('A timeout is not a reason to submit again').

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?

With annotations present, the description still adds substantial behavior: the return is a per-turn job_id consumed via read, long answers are persisted to answer_path/code_artifacts rather than returned inline, and external responses are explicitly flagged as data, not instructions. The idempotency guidance around timeouts and request_key reuse reinforces (and does not contradict) idempotentHint=true.

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 purpose, then constraints, then return/retrieval behavior; every sentence carries actionable content. It is dense and slightly repetitive around the saved-file behavior, keeping it short of ideal.

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 write tool with no output schema, it covers the return contract, persistence side effects, environment restriction, and retry semantics well. The remaining gap is the undocumented prompt/prompt_file distinction, which matters for correct invocation.

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 50%: request_key has a schema description and prompt_file does too, while prompt and thread_id do not. The description reinforces thread_id ('idle thread') and request_key ('unique key'), but says nothing about prompt vs prompt_file or how to choose between them, so it only partially compensates for the coverage gap.

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 concrete verb+resource ('Send one authorized follow-up in an existing personal conversation') and scopes it ('Personal chat only; never Custom Agent'). It references the read sibling for retrieving results, but never contrasts the tool with notion_personal_ask or notion_personal_submit, so the sibling differentiation is only partial.

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 ('Requires an idle thread and a unique request_key') and a recovery rule ('A timeout is not a reason to submit again'), plus routing to read for the correction. It stops short of the explicit when-to-use-this-vs-ask/submit decision the sibling set implies.

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: it discloses return identifiers (job_id, thread_id), states the idempotent dedup behavior that the idempotentHint only implies, warns that external responses are data rather than instructions (prompt-injection hygiene), and explains that long answers land in answer_path/code_artifacts rather than the tool output. This is exactly the kind of behavioral context annotations cannot carry.

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 first, then output behavior, then safety and retry rules. Every sentence carries distinct information with no padding, though the telegraphic staccato style makes the stream of constraints slightly harder to scan than a grouped structure would be.

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 fully carries the return-value burden: it names both durable IDs, explains where full answers are persisted, and advises reading the saved file for complete content. Combined with idempotency, timeout, and injection guidance, an agent has everything needed to call this 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?

Schema coverage is 67%, with prompt completely undocumented. The description adds meaning for request_key by explaining the dedup/reuse semantics, but it never explains the prompt vs. prompt_file choice, which is the schema's central oneOf constraint. Partial compensation for a real coverage gap.

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: 'Submit one authorized text task,' and scopes it sharply with 'Personal chat only; never Custom Agent.' It does not, however, name its siblings (notion_personal_ask/read/reply) or contrast its purpose against them, so the agent must infer the boundary rather than being told.

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?

Offers concrete operational guidance that few descriptions include: reuse request_key after errors, do not resend after a timeout, and never submit to a Custom Agent. These are clear when/when-not conditions, but they concern retry behavior only — there is no explicit guidance on when to pick this tool over the sibling ask/read/reply tools.

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