MCP server for Tenderly transaction simulation and trace debugging on EVM chains
Four tools with complete input schemas and descriptions. Tool names follow verb_noun pattern (simulate_*, get_*, list_*). Descriptions are clear and domain-specific (10-120 chars, within baseline). All parameters have types and descriptions. However, output schemas are not documented in the source code, LLMs cannot see what fields to expect from responses. Parameter descriptions lack format constraints (e.g., 'hex' for addresses, 'wei' for values). No error handling guidance visible. No tool annotations (readOnlyHint, etc.) despite all tools being READ_ONLY. Missing pagination details for list_simulations despite accepting limit/offset.
Retrieve a previously saved simulation by ID from Tenderly
List saved simulations from the configured Tenderly project
Simulate multiple EVM transactions in sequence against forked state
Simulate a single EVM transaction against forked state at a specific block
Output schemas not documented. LLMs cannot see what fields responses contain, forcing them to guess downstream field names and risk broken tool chains.
Tool annotations missing. All four tools are READ_ONLY but lack readOnlyHint in schema. Agents cannot distinguish safe tools from destructive ones without explicit hints.
Parameter format constraints missing. Descriptions say 'hex' and 'wei' but lack regex patterns or explicit format declarations. LLMs cannot validate input format before calling.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 52 | <=2025-11-25 | v2 |
list_simulations pagination incomplete. Accepts limit/offset but no documentation of total count, next_cursor, or result structure. Large result sets could exhaust context.
No error recovery guidance. Tool descriptions do not explain what to do on revert, API failure, or invalid input. Agents have no recovery path.