MCP server that manages, validates, and serves bank-specific ISO 20022 clearing profiles and rule packs.
Server has 4 well-named tools with clear verb-noun patterns (list_, get_, lint_, validate_). All tools have descriptions (avg 95 chars, within 10-1024 range). Input schemas are present and typed. However, descriptions lack depth on WHEN to use each tool and dependencies between them. Parameter descriptions are minimal (e.g., 'The profile to fetch' vs actionable guidance). Output schemas are not documented, LLMs cannot predict response structure for chaining. Error handling is absent from visible code. Tool annotations (readOnlyHint) are declared but not visible in schema definitions.
Return one clearing profile in full, including its rule bodies. Args: profile_id: The clearing profile identifier.
Evaluate a payload against a clearing profile and return findings. Args: payload_content: The raw ISO 20022 message text. profile_id: The clearing profile to lint against.
List the available clearing profiles as lightweight summaries. Use this to discover the ``profile_id`` values the other tools accept.
Validate a bank-supplied profile / rule-pack definition (raw JSON). Args: definition_content: The candidate profile definition, as JSON text.
Output schemas not documented. LLMs cannot predict response structure (fields, types, pagination) for downstream tool chaining or result extraction.
Parameter descriptions lack actionable guidance. 'The profile to fetch' does not explain format, constraints, or how to discover valid profile_id values. Descriptions should reference discovery tools.
No error handling guidance visible. Tools do not document what errors can occur, whether they are retryable, or what the LLM should do next (e.g., 'profile_id not found, call list_profiles first').
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
Tool descriptions do not explain dependencies or workflow. When should lint_payload be called vs validate_profile_definition? What does each return? Descriptions lack WHEN/WHY context.
lint_payload accepts raw ISO 20022 payload as string but does not document expected format, size limits, or encoding. LLMs may pass invalid payloads without guidance.