Server has 4 tools with basic schemas and descriptions, but significant gaps limit production readiness. telegram_status has good diagnostic output but lacks input schema documentation. telegram_execute_code is a dangerous catch-all with minimal guardrails. telegram_send_command and telegram_send_message have reasonable parameter schemas but descriptions lack actionable guidance on when to use each vs. the other. No error recovery hints, no parameter constraints (enums), no output schema documentation. Naming is verb-noun compliant but descriptions average ~150 chars, acceptable but generic. No tool annotations (readOnlyHint/destructiveHint) despite clear risk levels.

• Add tool annotations: @mcp.tool(readOnlyHint=True) for telegram_status; @mcp.tool(destructiveHint=True) for telegram_execute_code; @mcp.tool(destructiveHint=True) for telegram_send_command and telegram_send_message. This signals safety to LLMs without requiring description parsing.
• Document output schemas for all tools. E.g., telegram_status returns {connected: bool, user: {id, first_name, username}, error?: string}. telegram_send_message returns {message_id: int, timestamp: string}. Include these in docstrings or a separate schema registry.
• Replace free-form parse_mode string with enum constraint: parse_mode: Literal['md', 'html'] = 'md'. Update description: 'Parse mode for message formatting: md (Markdown) or html (HTML). Defaults to md.'
• Add confirmation step for telegram_execute_code: require a 'confirm' boolean parameter (default False) or implement a two-step pattern (preview code → confirm → execute). Prevents accidental data loss.
• Enhance error messages with actionable next steps. E.g., 'Session expired. Call telegram_status to diagnose, then provide a new session via TELEGRAM_SESSION_PATH environment variable.' Include links to setup docs.