Fact-check a claimed dollar amount about tax law ('the 2026 MFJ standard deduction is $32,200', 'CTC is $2,000 per child') against the corpus. Returns verified / refuted (with the correct value and citation) / unknown. Never states a verdict it cannot ground.
| Name | Required | Description | Default |
|---|---|---|---|
| asOf | No | ||
| query | Yes | what the amount is, e.g. 'standard deduction' | |
| filingStatus | No | ||
| claimedAmount | Yes | dollars, e.g. 50000 or "1234.56" |
Changes observed during successful MCP inspections. Dates show when Glama detected each change.
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It clearly states return values (verified/refuted with correct value and citation, unknown) and the grounding constraint, providing good transparency.
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?
Two sentences with no wasted words. Front-loaded with verb and resource, then return types and behavioral rule. Efficient and to the point.
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?
Tool has 4 params, no output schema. Description explains purpose, output types, and a behavioral rule. Missing details on optional parameters (asOf, filingStatus) and how they affect results, leaving some gaps.
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 coverage is 50% (2 of 4 params have descriptions). The description adds no extra meaning to parameters like 'asOf' or 'filingStatus' beyond examples. It does not compensate for missing schema descriptions.
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 clearly states the tool fact-checks a claimed dollar amount about tax law, with examples. It distinguishes from sibling 'verify_tax_claim' but not explicitly, and is specific about the resource.
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 implies usage for fact-checking dollar amounts but lacks explicit when-to-use or alternatives. It includes a behavioral rule ('Never states a verdict it cannot ground') but no guidance on when not to use.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
The compute tools are well-routed — calculate_tax explicitly directs business entities and fiduciaries to their own calculators — but verify_fact and verify_tax_claim are near-duplicates that differ only in whether the number being checked is a law parameter or a computed tax amount, and calculate_tax versus compute_return both accept the same individual-return input, distinguished only by output granularity. These two pairs create real misselection risk despite the detailed descriptions.
Every tool uses a snake_case verb-first pattern: calculate_*, compute_*, compare_*, determine_*, explain_*, find_*, list_*, lookup_*, search_*, verify_*. The single outlier, is_tipped_occupation, still follows the verb-first snake_case convention, so there is no mixing of styles.
Fifteen tools is at the top of the well-scoped band, but the domain is enormous — federal individual, business, and fiduciary computation, an 18-state return composer, and a citation-backed tax-law corpus with search, explanation, and verification utilities. Each tool has a distinct job, and the count feels lean rather than bloated for the stated scope.
The computational lifecycle is well covered: individual liability, the full 1040 line set, entity and fiduciary variants, state returns, filing-status comparison, dependent determination, and self-verification of any produced number. Minor gaps remain — only 18 of 50 states are modeled, and transfer taxes (Forms 706/709), payroll/FICA, and international forms are absent — but the tools refuse loudly on unmodeled territory rather than failing silently.
