Drillable

Verify claims against sources

verify

Read-only

Check claims against the corpora. SINGLE: verify({claim}) grounds one claim (core/verify truth-referee verdict where a domain surfaces it, else a corpus-grounding badge). RECEIPT: verify({claims:[…]}) claim-CHECKS each assertion and returns a TYPED per-claim verdict (verified / corrected / abstained) + drilled value + source, as a structured receipt (rides the response → portable to any client). Each claim is graded one of two ways: RECOGNITION-FREE — give op + params + asserted (STATE the computation you ran; it is RE-EXECUTED with no catalog match, so it grades NOVEL inputs too — the strongest form), or CATALOG-MATCH — give claim + asserted (matched to a covered record). Pass asserted (your own answer) to have it GRADED, not just grounded — the receipt corrects a confidently-wrong claim the model can't catch itself. A claim with no receipt line is unverified.

Input Schema

TableJSON Schema

Name	Required	Description
`op`	No	Recognition-free single check: the computation you performed (e.g. 'compare_decimal'). Discover ops via the recall-traps check_claim verb (no args).
`claim`	No	A single claim to check (back-compat / catalog-match). For several, use `claims`.
`claims`	No	Receipt mode: the claims to check. Each item is {claim?, op?, params?, asserted?, domain?} (or a bare string).
`domain`	No	Optional: restrict to one domain.
`params`	No	Recognition-free single check: inputs for `op`.
`asserted`	No	With a single `claim`/`op`: your answer, graded against the re-executed truth.

TDQS

A3.7/5.0

Behavior4/5

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

Annotations already indicate readOnlyHint=true, and the description adds significant behavioral context: two modes of operation, the distinction between recognition-free and catalog-match grading, and the meaning of 'asserted'. No contradictions with annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with the purpose but is somewhat verbose and dense, mixing examples and technical details. It could be restructured for clarity without losing information.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the complexity (6 params, nested objects, no output schema), the description covers parameter modes, return behavior (receipt vs badge), and grading logic, making it sufficiently complete for an agent to use effectively.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, baseline 3, but the description adds meaning beyond schema: explains how 'op' is discovered, how 'asserted' enables grading, and the relationship between parameters. This extra context justifies a 4.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool checks claims against corpora and describes two distinct modes (SINGLE and RECEIPT). However, it does not explicitly differentiate from sibling tools like 'search' or 'lookup', which could also involve fact-checking.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides internal guidance on when to use SINGLE vs RECEIPT mode and explains parameter combinations. However, it lacks explicit comparison to sibling tools or scenarios where this tool should not be used.

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

All Drillable tools Inspect

Try in Browser

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A4.3/5.0

Disambiguation4/5

Each tool has a distinct primary role—describe discovers, search/lookup/enumerate retrieve different shapes, compute aggregates, drill fetches addresses, verify checks claims, and notify_when_held tracks gaps. The main ambiguity is that compute and drill can both serve count/aggregate questions, though the descriptions do cross-reference each other to guide selection.

Naming Consistency4/5

Eight of the nine tools follow a clear lowercase single-word verb pattern: call, compute, describe, drill, enumerate, lookup, search, and verify. notify_when_held breaks that pattern with snake_case and a multi-word verb phrase, creating a minor but noticeable inconsistency.

Tool Count5/5

Nine tools is a well-scoped count for a knowledge retrieval and verification server. Each tool earns its place and the set avoids both bloat and thinness.

Completeness5/5

The tool surface covers the full intended lifecycle: discovery, search, glossary lookup, complete-set enumeration, aggregation, address-based drilling, claim verification, and registering unanswered questions. Abstentions and gap logging are built into misses, and describe/call unlock domain-specific verbs, so there are no obvious dead ends.

Resources

Need Help?