Three tools with basic descriptions but significant schema and parameter documentation gaps. python_execute and bash_execute have acceptable descriptions (100+ chars) but lack output schema documentation. Parameters use Annotated hints but lack formal JSON Schema type declarations in visible definitions. get_python_modules returns list[str] but no pagination or result limits documented. No error handling guidance, no parameter constraints (enums, ranges), and no security considerations for code execution tools. Tool names are verb-based (good), but descriptions do not explain WHEN to use each tool vs alternatives or what happens on failure.

• Add explicit output schema documentation to all tools. E.g., python_execute returns {type: 'string', description: 'stdout from Python execution, or error message on failure'}. Document what happens on timeout, syntax error, or permission denial.
• Add parameter constraints: timeout must be 1 - 3600 seconds (default 300); env dict keys must match /^[A-Z_][A-Z0-9_]*$/; code must be non-empty and <100KB.
• Enhance tool descriptions with decision guidance: 'Use python_execute for data processing, math, or library calls. Use bash_execute for system commands, file operations, or shell scripts. Both run in isolated containers.'
• Add error recovery guidance: 'If execution times out, try reducing code complexity or increasing timeout. If import fails, call get_python_modules first to check available packages.'
• Add pagination to get_python_modules: accept optional limit (default 50, max 500) and offset parameters; return {modules: [...], total: N, has_more: bool}.