ContextForge provides 8 tools with mostly complete schemas and descriptions. Naming follows verb_noun conventions (contextforge_retrieve, contextforge_query_graph, contextforge_patch_ast, memory_*). Descriptions are present and substantive (100-300 chars typical). However, several tools lack output schema documentation, parameter descriptions are sometimes generic, and error handling guidance is minimal. The contextforge_patch_ast tool has complex conditional logic (operation enum determines which params are required) that is documented but could be clearer. Memory tools are well-described but lack output structure details.

• Document output schemas for all 8 tools. For contextforge_query_graph, specify the structure of results (e.g., {symbols: [{name, type, location}], ...}). For memory_search, specify {results: [{id, content, importance, timestamp}], total_count}.
• Add error handling guidance to tool descriptions. E.g., contextforge_query_graph: 'If no results found, try a broader query_type (e.g., find instead of find_symbol).' memory_search: 'If no matches, try shorter keywords or check memory_list() to see available entries.'
• For contextforge_patch_ast, add a dry_run boolean parameter (default false) to preview changes without applying them. This prevents accidental code deletion.
• Clarify contextforge_patch_ast parameter dependencies in the schema using JSON Schema conditional constraints (if operation=replace_string then require search_string and replacement_string). Current description is clear but schema validation is weak.
• Add pagination support to memory_list and contextforge_query_graph results. Specify max result limits (e.g., memory_list returns max 50 entries; contextforge_query_graph returns max 20 symbols per query_type).