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

With no annotations provided, the description carries the full burden, and it delivers substantial disclosure: batching limits, the ad_group/adset alias behavior, single-use opaque continuation semantics, the min_as_of freshness-bound constraint, the tenant-owned ad_account_id requirement, and the bounded acknowledgement deadline with dependency_unavailable backoff. This is well beyond typical disclosure for a no-annotation tool. It stops short of stating read-only safety explicitly, though 'insight read' implies it, and the return shape ('evidence envelope') is left underspecified.

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 front-loaded with its core purpose and every sentence adds distinct information, so it is structured rather than padded. However, it is dense with domain jargon — 'Canonical Agent Method Profile V1,' 'common-wire adapter alias,' 'opaque single-use continuation,' 'dependency_unavailable backoff' — which raises cognitive load. It could convey the same guidance in plainer language, and the jargon detracts from readability without adding precision an agent needs.

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 16-parameter tool with no annotations and no output schema, the description covers pagination, batching, alias behavior, and the wait deadline well. But it leaves meaningful gaps: the consistency enum (cached/fresh_if_available/require_fresh) is never explained, the 'evidence envelope' return format is vague, and parameters like require_complete_range, after_mutation_ref, and date_range_mode get no contextual guidance. The coverage is good for the pagination core but incomplete for the full parameter 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 description coverage is only 6%, so the description must compensate, and it adds real meaning for several high-stakes parameters: scopes/scopes batching (up to 20, one tenant-owned ad_account_id each), min_as_of (legacy lower bound, never an anchor), wait_ms (deadline behavior), and group_by (the adset adapter-alias nuance). But with 16 parameters, many remain unelaborated — consistency enum semantics, fields, date_range_mode, require_complete_range, after_mutation_ref, response_mode — leaving the low-coverage schema only partially compensated.

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 clear verb-resource pair ('insight read') with concrete scope details: accepts one scope or a server-side batch of up to 20 scopes and returns 'one ordered compact evidence envelope.' This distinguishes it functionally from the continuation tool. However, it never explicitly differentiates itself from sibling readers like insights_query_rows, insights_query_overview, or insights_query_batch_overview, and the opening phrase 'Canonical Agent Method Profile V1' is opaque jargon that adds noise without clarity.

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 real usage guidance: start at page one, follow the server-returned continuation serially when has_more=true, and warns that min_as_of must never be used as a pagination anchor. It also specifies the wait_ms=0 fresh-request deadline with task_ref polling. But it never names alternatives or states when-not-to-use, which matters here given the sibling insights_query_consistent_continue exists for continuation — the description implies continuation but never tells the agent to route subsequent pages to that tool.

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