Server has 5 tools with basic schemas and descriptions, but significant gaps limit production readiness. Tool names follow verb_noun convention (good), but descriptions are minimal (avg ~50 chars, well below 194-char baseline). Parameters lack type specificity and validation guidance. No output schemas documented. Error handling returns plain strings without recovery guidance. Slack tools mix domain concerns (list_channels, send_message, get_messages) without clear composition. Weather tools are unrelated to Slack domain, suggesting scope creep. No tool annotations (readOnlyHint/destructiveHint). Secrets (bot_token) are passed as parameters rather than injected server-side.

• Inject bot_token server-side via environment variables (e.g. SLACK_BOT_TOKEN). Remove from tool parameters. Update tool signatures to not accept credentials.
• Expand all descriptions to 100-150 chars following pattern: 'WHAT: [action]. WHEN: [use case]. RETURNS: [structure].' E.g. 'List all non-archived channels in workspace. Call first to discover available channels before sending messages. Returns array of {id, name, topic, member_count}.'
• Document output schemas explicitly. For slack_list_channels, return structured JSON: {channels: [{id: string, name: string, topic: string, member_count: number}], total: number}. For slack_send_message, return {success: boolean, message_ts: string, channel_id: string}.
• Add tool annotations: mark slack_list_channels and slack_get_messages with readOnlyHint=true; mark slack_send_message with destructiveHint=true and idempotentHint=false to signal irreversibility.
• Implement error categorization: return {error: string, category: 'retryable'|'user_fixable'|'fatal', suggestion: string}. E.g. 'User not found. Try search_users() with partial name.' instead of bare error codes.

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

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