Loading ToolBench
A fast file search server using the plocate backend, with MCP (Model Context Protocol) support for integration with AI assistants
plocate-server defines 6 tools with clear naming (search, glob, fuzzy, reindex, stats, health) and reasonable descriptions (avg ~100 chars). All tools have input schemas with type definitions. However, parameter descriptions are minimal or absent in several tools, and output schemas are not documented. The reindex tool lacks error guidance for the 202 case. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles. Naming follows verb_noun convention well; composition is sound (each tool does one thing). Missing: output schema documentation, comprehensive parameter descriptions, error recovery guidance.
Search for files using multi-keyword ranked fuzzy matching. Returns matching file paths ranked by relevance.
Search for files using explicit glob patterns. Returns matching file paths from the plocate database.
Get server health status including instance name and base path configuration.
Trigger a background updatedb run to refresh the plocate index. Returns 200 when a new run begins, or 202 when one is already in flight.
Search for files using substring or glob patterns. Returns matching file paths from the plocate database.
Get server statistics including process metrics, index information, and reindex history.
Output schemas not documented. LLMs cannot infer what fields to expect from search, glob, fuzzy, stats, or health responses. This forces agents to guess at response structure and risks failed downstream tool calls.
Tool annotations missing. search, glob, fuzzy, stats, health are READ_ONLY; reindex is WRITE. No readOnlyHint or destructiveHint in tool definitions. Agents cannot distinguish safe-to-retry from irreversible operations.
reindex returns 202 (already-running) with no error guidance. Description mentions the status code but does not explain what the agent should do: retry later? Check stats? Abort the plan?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |
Parameter descriptions minimal or absent. 'q' in search/fuzzy and 'pattern' in glob have brief descriptions but lack format guidance (e.g., 'glob pattern must contain * ? [ ] to trigger glob mode'). stats and health have no input parameters documented.
No pagination or result limits documented. search, glob, fuzzy may return large result sets. No mention of max results, pagination support, or how agents should handle thousands of matches.