MCP Server integrations for ETH security workstreams
This server has 5 tools with basic schemas and descriptions, but significant gaps prevent higher scoring. Tool names follow verb_noun patterns (e.g., run_transaction, get_transactions_by_address), which is correct. However, descriptions are present but minimal (under 100 chars for most), parameter descriptions vary in quality, and no output schemas are documented. Most critically, there is no error handling guidance, no parameter validation rules in descriptions, and the schema quality varies significantly across tools. The server uses STDIO transport, which is a hard cap at 50 for protocolReadiness. For definitionQuality, the average across tools is 42, with major gaps in schema completeness, output documentation, and error guidance.
Retrieve all token activity from the provided address using the Sim API
Retrieve all transactions from the provided address using the Sim API
Retrieve function signature(s) for a given selector from 4byte
Retrieve source code for a given contract address
Re-run a transaction given its hash, returns a decoded trace
No output schemas documented. Tools return results but consumers cannot know what fields to expect, forcing LLMs to parse responses blindly. This violates the response-shaper pattern and increases hallucination risk.
Parameter descriptions lack constraint information. 'chain' and 'network' parameters in run_transaction have no enum values, format patterns, or examples of valid inputs. LLMs will guess or hallucinate values.
Error handling is missing. No tool provides recovery guidance, actionable error messages, or error classification. A failed API call will return a generic error with no context for the LLM to self-correct or retry intelligently.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 60 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 33 | - | v1 |
Descriptions are under 100 characters and lack context about WHEN to use each tool and WHAT it returns. 'Retrieve function signature(s) for a given selector from 4byte' tells the LLM what it does but not when to call it vs other tools or what the output looks like.
No pagination support documented. Tools like get_transactions_by_address and get_activity_by_address return lists but lack limit, offset, or cursor parameters. This risks context window exhaustion and violates the paginated-result pattern.
API key is injected from environment (ALCHEMY_API_KEY), which is correct for secret-injection. However, no documentation in tool descriptions mentions this requirement or how the tool handles missing credentials.