Audit a spreadsheet for formula and model errors

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

Annotations already establish readOnly=true, idempotent=true and closed-world, so the safety profile is covered. The description adds genuinely useful behavioral context beyond that: it performs an independent recalculation, it is heuristic rather than conclusive, and it points at likely fault locations rather than proving correctness. It stops short of describing result shape or severity-level meaning.

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 check list is front-loaded and every clause names a distinct defect class, so the length is earned rather than padding. The closing heuristic caveat is a single tight sentence that adds honest framing; structure is efficient overall.

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 readOnly annotations covering safety but no output schema, the description should carry more of the result contract. It doesn't say what a finding looks like, how severity is assigned or what 'limit' truncates. The heuristic caveat is a meaningful inclusion, but the undocumented severity/limit parameters leave the invocation picture incomplete.

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 50%: 'path' and 'recalc' are documented in the schema, but 'limit' and 'min_severity' have no descriptions anywhere. The description's mention of 'independent recalculation' merely restates the recalc param and it says nothing about the result-count limit or what the severity thresholds mean, so it does not compensate 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?

Opens with a specific verb+resource ('Audits an .xlsx workbook') and then enumerates the exact classes of defect it targets — error values, typed-over formulas, pattern-breaking formulas, hard-coded numbers, approximate-match lookups, volatile functions, external links, hidden sheets/rows/columns, broken named ranges. This scope is unmistakably distinct from the cell-level sibling inspect_cell, so an agent can route correctly 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?

The description explains the fitting context (when you suspect model errors) and honestly scopes the tool's power with 'it cannot prove a model is right'. However, it never names inspect_cell or any alternative, and gives no explicit when-not to use it, leaving the choice between workbook-level auditing and cell inspection to inference.

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