Bodhiorchard exposes 7 tools with significant definition gaps. Tool names lack consistent verb-noun structure (e.g., 'Bash', 'Read', 'Edit', 'Write' are bare nouns). Descriptions are present but minimal (10-50 chars), providing insufficient context for LLM selection. Input schemas are visible but sparse: most tools have only 1-2 parameters with basic type info. Critical issue: tools 4-7 (Bash, Read, Edit, Write) are inferred from guard service code, not explicitly registered in the MCP server definition, capping their scores at 50. No output schemas documented. Error handling is absent, no recovery guidance or actionable error messages. Parameter descriptions lack format constraints, ranges, or validation rules. The server relies on PreToolUse hooks for security rather than tool-level permission gates.

• Rename tools to verb_noun pattern: execute_bash_command, read_file, edit_file, write_file, search_bugs_in_codebase, check_feature_exists_in_codebase, update_bud_document.
• Expand descriptions to 100-200 chars. Example: 'search_bugs' → 'Search the codebase for potential bugs matching a query. Returns matching file paths, line numbers, and code snippets. Use this to identify issues before running checks.'
• Document output schemas for all tools. Example: Bash returns {status: int, stdout: string, stderr: string, execution_time_ms: int}.
• Add parameter constraints in descriptions. Example: 'command: Bash command to execute (max 500 chars, no pipes to external services, no rm/dd/mkfs allowed).'
• Explicitly register all 7 tools in server.py with full schema definitions. Verify Bash/Read/Edit/Write are not inferred but directly declared with input/output schemas.