Loading ToolBench
AI Runtime Operating System for Spring Boot applications with MCP (Model Context Protocol) integration for diagnostic and remediation tools.
Spring Lens MCP server exposes 12 tools with complete input schemas and descriptions, but quality is inconsistent. All tools have descriptions (10-200 chars, within baseline), and all parameters are typed with descriptions. However, descriptions are often generic and lack actionable context for LLM selection. Output schemas are not documented, critical for agent planning. Tool naming follows verb_noun convention well (list_, get_, query_, invoke_, apply_, promote_, draft_). No error handling guidance visible. Security concerns: invoke_runtime_tool and apply_overlay_instrumentation accept arbitrary JSON objects without validation hints. Missing pagination guidance on list tools. Overall: functional but below production baseline for agent reliability.
Register a new agent overlay specification.
Analyze an execution graph and return a structured diagnosis.
Generate overlay and patch drafts from runtime safety remediation findings.
Fetch a diagnostic playbook by problem id.
Query recent exception contexts from a Spring Lens runtime.
Fetch a structured execution graph by execution id.
Query recent slow SQL observations from a Spring Lens runtime.
Output schemas not documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., what fields does get_slow_sql return? Does it include executionId for chaining to diagnose_execution_graph?). Baseline: 100% of A+ tools document return types.
invoke_runtime_tool and promote_runtime_safety_remediation accept 'arguments' as untyped object with no validation hints. LLMs cannot determine valid argument structure, required fields, or constraints. Baseline: all parameters need descriptions explaining format and constraints.
No error handling guidance. Tools lack recovery hints (e.g., 'If applicationId not found, call list_registered_apps first'). Baseline: error responses must tell LLM what to do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2025-06-18+ | v2 |
Invoke a capability-contributed runtime tool on a registered application.
List registered Spring Lens runtime applications.
List capability-contributed runtime tools from a registered application.
Submit runtime safety overlay drafts and patch drafts into the control plane.
Query captured probe values from a registered application.
Descriptions lack actionable context. E.g., 'Query recent slow SQL observations' does not explain when to use this vs diagnose_execution_graph, what 'recent' means, or what minDurationMs default is. Baseline: descriptions should state WHAT, WHEN, and prerequisites.
No pagination guidance on list tools. list_registered_apps and list_runtime_tools lack limit/offset parameters and total count in response. Baseline: tools returning lists must support pagination to avoid context window exhaustion.