A cryptocurrency implementation based on Bitcoin, featuring multi-algorithm mining support and various blockchain extensions
This MCP server exposes 2 blockchain tools with minimal documentation. Both tools have very short descriptions (under 20 characters when stripped of context), lack comprehensive parameter documentation, and provide no documented output schemas. The server appears to wrap SmileyCoin RPC calls with minimal abstraction. Tool names follow a basic verb_noun pattern (getblock, getblockhash) but descriptions are severely underdeveloped. No evidence of error handling guidance, input validation documentation, or output schema definition. The codebase shown is primarily build infrastructure (Dockerfile, Makefiles, Python setup) with no explicit MCP server implementation visible, tool definitions appear inferred from RPC function names rather than explicitly registered.
Retrieve block data by hash with optional verbosity
Get block hash by block height index
Tool descriptions are extremely brief (under 20 characters) and lack context about WHEN to use each tool or WHAT problem they solve. 'Retrieve block data by hash with optional verbosity' and 'Get block hash by block height index' lack guidance on return types, use cases, or dependencies.
No documented output schemas visible. LLMs cannot infer what fields are returned by getblock() or getblockhash(). This violates the requirement that tools document their return structure so agents can plan downstream calls and extract the right data.
Parameter descriptions are minimal. 'Block hash' and 'Block height' do not explain format (hex string? length? case sensitivity?), valid ranges, or examples. The 'verbose' boolean in getblock lacks guidance on what fields differ when true vs false.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 42 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 0 | - | v1 |
No error handling documentation. What happens if a block hash is invalid? What if a height index is out of range? Are errors retryable? Should the agent call a discovery tool first? No guidance provided.
Tool definitions appear inferred from blockchain RPC function names rather than explicitly registered in visible MCP server code. No source code showing MCP tool registration, schema declarations, or server initialization is provided. This caps per-tool scores at 50 per the hard rules.
No evidence of pagination, limiting, or result shaping. If getblock returns verbose blockchain metadata (raw hex, difficulty, nonce, etc.), this wastes tokens and confuses LLMs. No indication that responses are stripped to essential fields only.
Tool names are adequate (verb_noun pattern) but lack disambiguation guidance. If future tools like 'getblockhash', 'getblockhashes', 'searchblocks', or 'listblocks' are added, naming conflicts could confuse LLMs about which to select.