Loading ToolBench
Local-first continuity layer for AI work across tools, providers, and fresh sessions. Provides MCP tools for searching skills, managing handoffs, and syncing context across AI coding assistants.
Context Engine provides 6 tools with complete input schemas and descriptions. Naming follows verb_noun convention (search, list, get, sync). Descriptions are adequate (80-180 chars) but lack LLM-optimized guidance on when to use each tool vs. alternatives. All tools have typed parameters with descriptions. Output schemas are not explicitly documented in the source, responses are inferred from API calls. Error handling is minimal; no recovery guidance or categorization. The server is well-structured but lacks production-grade polish around error messaging and output documentation.
Retrieve detailed information about a specific skill, including its sections, body content, and metadata.
List handoff documents for resuming work across sessions. Can filter by repository or thread tag. Returns handoff metadata including staleness and commit timeline.
List all available skills with their metadata, active status, and trigger information. Can filter by category or active status.
Search across indexed skills and context using semantic search. Returns ranked results with skill metadata and relevance scores.
Get the current status of the Context Engine index, including chunk count, skill count, embedding model, and readiness state.
Synchronize or create a handoff document for a project repository. Returns the created or updated handoff metadata.
Output schemas not documented. LLMs cannot plan downstream calls or extract fields without knowing response structure. Responses are inferred from HTTP calls but not formally declared.
Error handling lacks recovery guidance. No categorization of errors as retryable, user-fixable, or fatal. CEUnreachableError and CEHttpError exist but responses do not guide LLM next steps.
Descriptions lack LLM-optimized guidance on when to use each tool. E.g., context_engine_search vs. context_engine_list_skills distinction is unclear. No dependency hints or prerequisites stated.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | <=2025-11-25 | v2 |
context_engine_sync_project_handoff is a WRITE operation but description does not explicitly state it modifies state. Agents need to know which calls are safe to retry.
No pagination support visible in list tools (context_engine_list_skills, context_engine_handoffs). Large result sets could exhaust context window without limit/offset parameters.