scan_topic

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

With no annotations, the description carries the full burden and does so thoroughly: it discloses the returned buckets (candidates, digest, images, coverage, degraded/ok/cost), that degradation signals are surfaced, that a sample-bias declaration is included, that digest images are low-resolution 160px thumbnails unfit for reading numbers, and how ranking changes when families are supplied. Defaults come with rationale (e.g. 12 comment budget, 15 comments per candidate).

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 one-line purpose and call-replacement claim, then organized into return-value, parameter, offset, exclusion and ranking sections. It is long and occasionally restates what the output schema already defines, plus it leans on internal jargon (批次 D2/D3) — but for a 14-parameter orchestration tool nearly every line carries decision-relevant 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?

Covers the full lifecycle an agent needs: what the call does, how to tune cost versus coverage, how to paginate via coverage.next_offset and rest_preview, how to avoid re-scanning already-analyzed videos, and how the ranking differs with and without families. Even though an output schema exists, the extra explanation of degraded/cost/coverage makes the tool safe to invoke blind.

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 14 parameters, so the description must compensate, and it does: it explains keywords (4–6, five semantic categories), per_query, comment_budget, comment_n, image_policy with its three values, image_top, families with its structure and consequence (kw_families returns null if omitted), pages_per_keyword, min_play, published_after format, offset, exclude_topic and exclude_bvids. Only max_candidates goes unmentioned.

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 composite action — multi-keyword search → dedupe → rank → pull comments for top candidates — and explicitly positions itself as the single entry point for broad Phase-1 scanning, replacing N searches + M metadata + M comment calls. This clearly separates it from siblings like search_videos, get_metadata and get_comments, which are single-step operations.

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 rich usage guidance: when to use offset (recall/return gap), when to use exclude_topic (second round on the same topic, stackable with exclude_bvids), and when a filter narrows vs. widens the pool. It does not name sibling tools or a Phase-2 counterpart as explicit alternatives, so the routing information is implied rather than stated.

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