Server has 4 tools with complete input schemas and descriptions. Naming follows verb_noun convention (flashcard_parse, flashcard_sync, flashcard_list, flashcard_check). Descriptions are present (avg 95 chars) but lack actionable context for LLM selection, they state WHAT but not WHEN or WHY to use each tool. Parameters have types and descriptions, but lack constraints (enums, ranges, patterns). No output schemas documented. Error handling returns JSON with 'error' field but lacks recovery guidance or categorization. Missing tool annotations (readOnlyHint, destructiveHint). No pagination support despite list tool returning potentially large results.

• Document output schemas for all tools. Example: flashcard_parse should return {cards: [{front: string, back: string, deck: string}], count: number, errors: [{file: string, reason: string}]}
• Add pagination to flashcard_list: accept offset/limit parameters, return total_count and next_offset. Cap results at 50 by default.
• Enhance tool descriptions with WHEN/WHY context. Example: 'Parse a single file to preview cards before syncing. Use this to validate formatting or check what will be created. For bulk operations, use flashcard_sync instead.'
• Add tool annotations: mark flashcard_parse and flashcard_list with readOnlyHint=true; mark flashcard_sync with destructiveHint=true and support dryRun=true for confirmation.
• Improve error responses: categorize as retryable (connection timeout), user-fixable (invalid file path), or fatal (vault not found). Include actionable guidance: 'Vault path not found. Check VAULT_PATH environment variable is set correctly.'