Loading ToolBench
MCP server providing Remembrance auto-query hooks, skill docs, and contribution tools for Claude Code, VS Code, Cursor, OpenClaw, and opencode plugins
Remembrance MCP server has 15 tools with basic schemas and descriptions, but significant quality gaps prevent production use. All tools have descriptions (10-80 chars, mostly under 100), but many are generic and lack actionable context. Input schemas are present but minimal, most parameters lack type constraints, enums, or validation guidance. No output schemas documented. Error handling is absent. Tool naming is inconsistent (query_skills vs invoke_skill vs submit_remembrance). No tool annotations (readOnlyHint/destructiveHint). Critical issue: descriptions are too brief to guide LLM selection (e.g., 'Query Remembrance for matching skills based on context' does not explain when to use this vs get_skill). Parameters like 'parameters' (object) in invoke_skill lack any structure documentation. No pagination, no field-level descriptions for complex inputs.
Retrieve a specific resource by slug from Remembrance
Retrieve a specific skill by slug from Remembrance
Invoke a Remembrance skill with parameters
Propose a private skill for an organization
Propose a new skill idea to Remembrance
Query Remembrance for matching skills based on context
No output schemas documented. LLMs cannot plan downstream calls or extract required fields (e.g., what does query_skills return? Is there a skill_id for get_skill?). Breaks tool chaining.
invoke_skill has 'parameters' as bare object with no structure documentation. LLM cannot infer what fields to pass. Violates pattern:constrained-input.
Descriptions are too brief (30-40 chars average) and lack context. 'Query Remembrance for matching skills based on context' does not explain when to use this vs get_skill, what 'context' means, or what the LLM should expect back. Violates pattern:tool-description.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
Retry submission of a private lesson candidate
Submit general feedback to Remembrance
Submit feedback on preference compatibility
Submit a candidate for a private lesson
Submit feedback on a Remembrance query result
Submit a new remembrance/lesson to Remembrance
Submit a resource to Remembrance
Submit a review of a Remembrance resource
Submit a suggestion to Remembrance
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). LLM cannot distinguish safe reads from destructive writes. All 8 submit_* tools are WRITE but unmarked.
No error handling guidance. Tools do not document what errors are retryable, user-fixable, or fatal. No recovery hints (e.g., 'If skill not found, try query_skills first'). Violates pattern:recovery-guide.