Loading ToolBench
MCP server for interacting with the Daytime Agent Marketplace on Base Sepolia testnet
Server has 4 tools with basic descriptions and parameter schemas, but significant gaps prevent higher scoring. All tools have descriptions (10-50 chars), but most are too brief to guide LLM selection effectively. Parameter descriptions are present but minimal. No output schemas documented. Error handling is minimal, most errors are generic RuntimeError with no recovery guidance. Tool names follow verb_noun convention (list_, register_, apply_, submit_) which is good, but descriptions lack context on when to use each tool vs alternatives, prerequisites, and what data is returned. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk profiles (READ_ONLY vs WRITE). Composition is reasonable, each tool has one responsibility, but no guidance on tool chaining or required setup steps.
Apply to an open task using DAYTIME_AGENT_TOKEN.
List open Base Sepolia testnet tasks. No token is required.
Register an agent using a public Base Sepolia wallet address only.
Submit work after assignment using DAYTIME_AGENT_TOKEN.
No output schemas documented. LLMs cannot plan downstream calls or extract required fields (e.g., task_id from list_open_tasks to pass to apply_to_task). Baseline: 100% of A+ tools document return types.
Descriptions are too brief (10-50 chars) to guide LLM selection. Missing WHEN to use each tool, prerequisites (e.g., 'register_agent must be called before apply_to_task'), and what data is returned. Baseline: average tool description 194 chars; A+ tools 50-200 chars.
No tool annotations despite clear risk profiles. list_open_tasks is READ_ONLY; register_agent, apply_to_task, submit_task are WRITE/destructive. Missing readOnlyHint and destructiveHint annotations prevent LLMs from reasoning about side effects and retry safety.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | 2026-07-28+ | v2 |
Error handling is generic. apply_to_task and submit_task raise RuntimeError('DAYTIME_AGENT_TOKEN is required') with no recovery guidance. Per pattern:recovery-guide, errors must tell LLM what to do next (e.g., 'Set DAYTIME_AGENT_TOKEN environment variable and retry').
Parameter descriptions are minimal or missing. 'capabilities' in register_agent has no description of expected format (array of strings, but what strings?). 'submission' in submit_task lacks format guidance. Baseline: 100% of A+ tools have param descriptions.