run_overview

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

With no annotations at all, the description carries the full burden and does substantial work: it discloses null semantics for open_items/answered_items on old service versions, that card is null only when the route is absent, that zero-data runs (documents, failed verification) return {cells:0, blank:0}, and that document runs return [] rather than null. That is unusually rich edge-case disclosure. It does not state read-only-ness or any permission/auth expectations, which keeps it from 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?

Purpose is front-loaded in the first clause, then the text moves systematically through open_items, answered_items, card, and the old-service/no-data fallbacks. Every sentence covers a distinct return-value case rather than padding. It is dense and semicolon-heavy, which slightly hurts scanability, but little is genuinely redundant.

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, so the description must explain return values — and it does, covering the main fields, their null/empty fallbacks, and the cross-reference to get_card. Given a one-parameter read tool, an agent has enough to call it and interpret the response. Minor gaps remain around what "稳定性" and the health checks actually measure, but those are domain concepts rather than calling prerequisites.

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 parameter saved_as is fully documented in the schema (100% coverage), including its provenance (from list_runs or build_ontology). The description adds nothing further about the parameter — no format, no validation, no aliasing behavior. Baseline 3 is appropriate when the schema already does the work.

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 opening clause names a specific resource and enumerates what the overview contains (object/relation counts, health-check pass count, answerable questions, stability, confirmation status), so an agent knows exactly what it gets back. It also implicitly separates itself from get_card ("看细节用 get_card"). It stops short of distinguishing itself from siblings like list_runs or health, which also touch run-level state.

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 real guidance for the open_items field ("先看它" — check this first, it is what you bring to the customer) and a pointer to get_card when cell-level detail is needed. However, there is no statement of when to call run_overview versus list_runs, health, or confirm_ontology, and no explicit exclusion cases. Usage is implied rather than prescribed.

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