Loading ToolBench
MCP JSON-RPC dispatcher for Aracne2 digital humanities platform. Exposes tools for collection browsing, document retrieval, entity search, and authority lookups over Streamable HTTP.
Aracne2 provides 8 read-only tools with clear, domain-specific descriptions (avg ~120 chars). All tools follow verb_noun naming (list_, get_, search_, find_, lookup_). However, input schemas are not visible in the provided source code, only tool names and descriptions are shown. Parameter types, constraints, and output schemas cannot be verified. Error handling guidance is absent. The server targets a specialized humanities/digital humanities use case (TEI XML, named entities, authority lookups) with coherent composition, but lack of visible schema definitions and error recovery patterns prevents a higher score.
List the document occurrences of one indexed entity, with the surrounding context. Filtered to the corpus.
Fetch a single collection by slug. Includes description, license id, schema id, target date, and document count.
Return the raw TEI XML of a single document. Truncated at 2 MB; the response carries a hint when truncation occurs.
List public, published collections visible to this token (scoped to the corpus). Returns slug, title, description, and publication date.
List the TEI document filenames inside a collection.
Resolve a name through an external authority API (Wikidata / ORCID / ROR / VIAF) and return canonical IDs + URIs. Useful when the assistant needs an authoritative URI for an entity not yet indexed in this instance.
Input schemas not visible in source code. Cannot verify parameter types, constraints, enums, or required/optional status for any of the 8 tools.
Output schemas not documented. Descriptions mention return fields (e.g., 'slug, title, description' for list_collections) but formal schema definitions are absent.
No error handling guidance. Tools lack recovery hints (e.g., 'If collection not found, try list_collections() first'). No categorization of retryable vs. fatal errors.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
Search the named-entities index. Returns canonical form, TEI tag type, authority URI, and occurrence count for the matches that appear inside the corpus.
Strip TEI markup from a document and return the body text. Lighter on the LLM context than get_document_source when the assistant only needs the prose content (no apparatus).
Pagination not mentioned. list_collections and list_documents likely return multiple items but no limit/offset/cursor parameters are documented.
Parameter descriptions missing. Tool descriptions mention what is returned but do not describe input parameters (e.g., what does 'slug' mean for get_collection? Is it required? Format constraints?).