Server has 10 tools with consistent naming (verb_noun pattern: tg_send, tg_read_dialog, tg_search_local_history, etc.). All tools have descriptions (50-150 chars, within baseline 34-392 range). Input schemas are present with typed parameters and descriptions. However, output schemas are not documented in the visible code, LLMs cannot predict return structure. Error handling is mentioned in docstrings (returns ERROR:/WARNING: text) but not exemplified. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications (WRITE vs READ_ONLY). Parameter descriptions are adequate but lack format constraints (e.g., 'target' accepts @username, phone, or ID but no regex/enum). No pagination support visible for list_dialogs or search results despite baseline expectation for discovery tools.

• Document output schema for each tool. Example: tg_send returns {success: bool, message_id: int, timestamp: string}. tg_list_dialogs returns {dialogs: [{id, name, type, unread_count}], total: int}.
• Add tool annotations: mark tg_send, tg_mark_read, tg_add_contact, tg_persona_set with destructiveHint=true; mark all READ_ONLY tools with readOnlyHint=true.
• Constrain 'target' parameter: add enum or pattern. E.g., pattern: '^(@[a-zA-Z0-9_]{5,32}|\+?[0-9]{10,15}|[0-9]{1,20})$' with description 'Username (@handle), phone (+1234567890), or numeric peer ID'.
• Add pagination to tg_list_dialogs and tg_search_local_history: include offset/limit in request, return total_count and next_offset in response.
• Enrich error messages with recovery guidance. E.g., 'Peer not found. Try tg_list_dialogs() to see available contacts' or 'Rate limited. Retry after 60 seconds.'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry