read_parquet_columns_tool

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

Beyond the readOnly/openWorld/non-destructive annotations, the description discloses non-obvious behavior: limits do not reduce full-file HTTPS transfer (a cost/latency warning), nested arrays count against an output budget, and no scientific filtering happens automatically. It stops short of covering auth, failure modes, or how the two addressing paths interact.

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 clauses, front-loaded with the core action and then the constraints; nothing is padded. Some phrasing is compressed to the point of ambiguity ("kernels", "output budget") without a definition.

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 output schema covers return values, so that omission is fine, and the transfer-cost and filtering caveats are valuable. But for an 11-parameter tool with zero schema descriptions, leaving s3_uri vs file_url, the partition filter semantics, and inv_mode/freq_mode entirely unaddressed is a substantive gap an agent could stumble on.

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 11 parameters, so the description carries a heavy burden. It does explain file_url (must be catalog-listed), columns (physical columns), the existence of storage partition filters, and that row/column limits exist, but s3_uri, project, product, max_files, inv_mode and freq_mode get no explanation at all, and the partition filter names are never mapped to year/month.

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?

"Read selected physical columns" gives a concrete verb and resource, and "physical" plus "preserving nulls, nanoseconds and kernels" signals a raw, columnar read. It does not name or contrast with read_parquet_preview_tool, so the agent must infer the difference between a full column read and a preview from the names alone.

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?

"Supply an exact catalog-listed file_url, or storage partition filters" gives real input-selection guidance and an important constraint (the URL must be catalog-listed). However, it never says when to prefer this tool over the preview tool or when to first call the listing tools, so usage context 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.