Loading ToolBench
MCP server for AI Agent Wallet — provides tools for wallet operations, transaction simulation, policy evaluation, audit logging, and pact management
Server has 9 tools with basic naming (verb_noun pattern) and descriptions present. However, descriptions are minimal (10-50 chars), parameter descriptions are sparse or missing, and output schemas are not documented. Input schemas exist for most tools but lack depth. No error handling guidance, no recovery patterns, and no tool annotations. Naming is clear but descriptions fall short of LLM-optimization standards (target 50-200 chars). This is a typical community server with functional but incomplete definitions.
Get the wallet address
Get the wallet balance
Get details of a specific pact
Get the current policy configuration
List task-scoped authorization pacts
List pending transactions awaiting approval
Propose a transaction for approval
Query the audit log
Simulate a transaction to assess risk and gas usage
Descriptions are too brief (10-50 chars). LLMs need 50-200 char descriptions explaining WHAT the tool does, WHEN to use it, and any prerequisites. Current descriptions lack context for tool selection.
Output schemas are not documented. LLMs cannot plan downstream calls or extract required fields without knowing what each tool returns. Document return types for all tools.
No error handling guidance. Tools lack recovery patterns (e.g., 'If transaction simulation fails, check gas limits' or 'If pact not found, call list_pacts() first'). Agents cannot self-correct on failures.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 38 | 2026-07-28+ | v2 |
Parameter descriptions are missing or minimal. 'limit' in query_audit lacks format/range guidance. 'status' in list_pacts lacks enum values. LLMs cannot infer valid inputs from names alone.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Agents cannot distinguish safe reads from destructive writes without explicit hints. propose_tx is marked WRITE but lacks annotation.