Naeth provides 10 tools with complete input schemas and descriptions in Spanish. Naming follows verb_noun convention (memory_add, memory_search, relation_add). Descriptions are moderately detailed (avg ~120 chars) but lack explicit guidance on when to use each tool vs. alternatives. Parameters are typed and described, but output schemas are not documented in the visible code. Error handling and recovery guidance are absent. The server implements a sophisticated memory/graph system but tool definitions do not expose enough structure for LLMs to reason about chaining or error recovery.

• Document output schemas for all tools. Specify the JSON structure returned by each tool, including field names, types, and whether results are paginated. Example: 'Returns {id: string, content: string, digest: string, created_at: ISO8601, is_current: boolean, versions: [{id, created_at}]}'.
• Add error handling guidance to descriptions. For memory_search, add: 'Returns empty array if no matches found. If query is malformed, returns 400 with message indicating valid query syntax.' For memory_get, add: 'Returns 404 if ID not found; check memory_search() to discover valid IDs.'
• Use tool annotations (idempotentHint, destructiveHint, readOnlyHint) instead of embedding this info in text. Mark memory_add with idempotentHint=true, memory_tombstone with destructiveHint=true, memory_search with readOnlyHint=true.
• Clarify tool selection in descriptions. For memory_supersede, add: 'Use this to create a new version that replaces an old one, keeping history. Use memory_add for new independent memories.' For memory_tombstone, add: 'Logical delete (reversible). The memory remains in the database but is marked deleted.'
• Translate descriptions to English. LLMs perform better with English descriptions. Provide both Spanish and English, or standardize on English for tool definitions.

Inferred effective spec: <=2025-11-25.

• Relies on Dynamic Client Registration (deprecated) - use Client ID Metadata Documents (CIMD)