Loading ToolBench
Model Context Protocol server adapter for the OpenHCS agent API, providing high-content screening image processing capabilities with GPU support
OpenHCS MCP server has 6 tools with basic schemas and descriptions, but significant quality gaps. All tools have descriptions (10-80 chars, mostly short), and input schemas are present with typed parameters. However, descriptions are terse and lack context for LLM selection. No output schemas documented. Parameters lack constraints (enums, ranges, format hints). Error handling guidance is absent. Tool names follow verb_noun pattern but lack clarity on distinctions (e.g., openhcs_debug_inspection vs openhcs_send_debug_command). No tool annotations (readOnlyHint/destructiveHint). Composition is reasonable, tools are single-purpose, but chaining support is unclear (no documented output fields). Security: timeout_ms and host/port parameters are exposed, but no secrets in params. Overall: functional but below production baseline (median ~50 for community servers; this scores 42 due to missing output schemas, sparse descriptions, and no error guidance).
Inspect paused debug worker runtime state
Get execution status from runtime server
Export artifact from paused debug worker
Scan for OpenHCS ZMQ runtime servers on specified host and ports
Send command to paused debug worker
Query runtime server information and capabilities
No output schemas documented. LLMs cannot infer what fields to expect from tool responses, forcing them to guess at downstream chaining and data extraction.
Descriptions are terse (10-80 chars) and lack context. Missing WHAT the tool does, WHEN to use it, and prerequisites. E.g., 'Query runtime server information and capabilities' does not explain what 'capabilities' means or when to call this vs openhcs_scan_runtime_servers.
Parameters lack constraints. 'transport_mode' accepts free-form strings (no enum); 'timeout_ms' is unbounded (no min/max). LLMs will hallucinate invalid values like transport_mode='QUIC' or timeout_ms=999999999.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | 2026-07-28+ | v2 |
No error handling guidance. Tools do not document what errors are retryable, user-fixable, or fatal. No recovery hints (e.g., 'If server not found, try openhcs_scan_runtime_servers first').
No tool annotations. Tools marked as READ_ONLY or WRITE lack readOnlyHint/destructiveHint/idempotentHint in schema. LLMs cannot infer safety properties and may retry destructive operations.