Konteks CLI exposes 6 tools with complete input schemas and descriptions. Naming follows verb_noun convention (recall, search, save_*, forget, warm_up). Descriptions are concise (20-80 chars) but lack depth on WHEN to use each tool and what downstream actions they enable. Parameter descriptions are present but minimal, most lack format constraints, ranges, or dependency hints. Output schemas are not documented in the source. Error handling and recovery guidance are absent. Tool composition is sound (each does one thing), but parameter design could better match the chat data model (e.g., 'focus' as free-form strings vs. structured queries).

• Document output schemas for all 6 tools. Example: 'konteks_recall returns {results: [{id, source, score, excerpt}], total_count}'. This enables agents to chain tools and extract required fields.
• Expand tool descriptions to 100-150 chars with WHEN/WHY guidance. Example: 'Recall compact, task-relevant project context before answering or working. Use this when you need a quick briefing on the current project. Use search_* for targeted queries on specific topics.'
• Add min/max bounds and format constraints to numeric and string parameters. Example: 'limit (1-100, default 20): Maximum number of results to return.' 'focus (array of 1-5 strings, each 5-50 chars): Natural-language focus areas for retrieval.'
• Clarify mutual exclusivity in konteks_forget. Example: 'id (string, optional): Memory ID to delete. Mutually exclusive with query. mode (enum: hard_delete|invalidate|soft_delete): Deletion strategy. hard_delete removes permanently; invalidate marks stale; soft_delete hides from results.'
• Add error handling and recovery guidance. Example: 'If konteks_forget returns "ID not found", call konteks_search() with a query to find the memory ID first. If konteks_save_memories rejects a memory, check that importance is 0-1 and kind is one of: fact, decision, pattern, warning.'