The Meta Ads MCP server has consistent naming patterns and clear parameter descriptions, but lacks documented output schemas, error handling guidance, and some parameter constraints. All 12 tools start with action verbs (get_, create_, update_, search_), which is a strong foundation. However, tool descriptions are concise (mostly 6-15 words), falling below the production baseline of 194 chars, they lack context about when to use each tool, dependencies, or recovery guidance. Parameter schemas are present with type declarations and descriptions, but output schemas are entirely undocumented, LLMs cannot plan downstream calls without knowing what fields to expect. Error handling is absent: there are no recovery guides, no distinction between retryable and fatal errors, and no actionable error messages. The server uses HTTP transport (fastmcp + uvicorn), which is current. Overall, this is a solid tier-C implementation with good naming discipline but significant gaps in description richness, output documentation, and error guidance that would cause agents to struggle in production.

• Document output schema for every tool. Example: 'Returns: {"accounts": [{"id": string, "name": string, "currency": string, "timezone": string}], "total_count": integer}'. This enables agents to extract fields for downstream calls.
• Expand tool descriptions to 50-150 characters. Instead of 'List campaigns for an ad account', write: 'List ad campaigns for an account. Returns campaign IDs, names, status, budget, and objective. Use to discover campaigns before updating or fetching performance metrics.'
• Add enum constraints to parameter descriptions. Change 'Campaign objective (e.g., OUTCOME_TRAFFIC, OUTCOME_AWARENESS)' to '(required) Campaign objective: one of OUTCOME_TRAFFIC, OUTCOME_AWARENESS, OUTCOME_ENGAGEMENT, OUTCOME_LEADS, OUTCOME_SALES, OUTCOME_APP_INSTALLS'.
• Add error recovery guidance for create_campaign and update_campaign. Example: 'Error response: {"error": "Invalid objective", "valid_values": ["OUTCOME_TRAFFIC", ...], "suggestion": "Use search_interests to identify targeting before creating campaign"}'. Return specific next steps.
• Implement dry-run support for write tools. Add an optional 'dry_run' boolean parameter to create_campaign and update_campaign. When true, validate but do not execute, returning a preview of changes.
• Document pagination semantics for list tools. Clarify whether get_campaigns returns a next_cursor or uses offset-based pagination, and whether it includes a total_count field.
• Add parameter format constraints to descriptions. For account_id: '(string, 8-16 digits or 'act_' prefix + 8-16 digits, e.g., 'act_123456789')'. For campaign_id and adset_id: '(string, numeric ID returned by get_campaigns or get_adsets)'.

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

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