Thrift provides 10 well-scoped, read-only tools with consistent naming (verb_noun pattern: workspace_info, find_files, grep_text, rank_files, file_window, compress_log, count_tokens, plan_context, classify_task, cost_report). All tools have descriptions (avg ~100 chars, within baseline 34 - 392). Input schemas are present with type declarations and parameter descriptions. However, output schemas are NOT documented in the source, LLMs cannot predict return structure. Error handling is minimal; no recovery guidance or actionable error messages visible. No tool annotations (readOnlyHint, etc.). STDIO transport caps protocol readiness at 50, but definition quality itself is moderate-to-good: naming is clear, descriptions are adequate, schemas are present but incomplete.

• Document output schema for each tool. E.g., workspace_info returns {file_count: int, language_distribution: {lang: count}, salience_diagnostics?: object}. Include field types and whether fields are optional.
• Add error handling guidance to each tool description. E.g., 'If workspace root is invalid, returns error code INVALID_ROOT with message suggesting valid paths. Retry with corrected root or ask user.'
• Declare tool annotations in schema. Add 'readOnlyHint': true to all 10 tools' input schemas to signal safety to LLMs.
• Enhance parameter descriptions with format/constraint details. E.g., 'pattern: fnmatch-style glob (e.g., *.py, src/**/*.js); radius: 1 - 50 lines (default 5); budget_tokens: 100 - 1000000 (default 8000).'
• Clarify cost_report 'since' enum: 'all = all recorded events; session = last THRIFT_SESSION_HOURS (default 4 hours, configurable via env).' Add example: 'since=session returns costs from the last 4 hours.'