MCP server for managing Google Ads and Facebook Ads campaigns with GAQL query execution, keyword planning, and account management
Three tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. run_gaql and run_keyword_planner have reasonable descriptions (100+ chars), but list_accounts lacks detail. Parameter descriptions are present but sparse (e.g., 'The Google Ads customer ID' without format guidance). No output schemas documented. Error handling is minimal, tools raise generic exceptions without recovery guidance. No tool annotations (readOnlyHint, etc.) despite all being read-only. Naming is acceptable (verb_noun pattern), but descriptions don't explain WHEN to use each tool vs. alternatives.
List all accessible Google Ads accounts including nested sub-accounts.
Execute GAQL using the non-streaming search endpoint for consistent JSON parsing.
Generate keyword ideas using Google Ads KeywordPlanIdeaService. This tool allows you to generate keyword ideas based on seed keywords or a page URL. You can specify targeting parameters such as language, location, and network to refine your keyword suggestions.
No output schemas documented. LLMs cannot predict response structure (fields, types, nesting). run_gaql returns {'totalRows', 'results'} but this is inferred, not declared. list_accounts returns {'accounts', 'total_accounts'}, structure of account objects is undocumented.
list_accounts description is too brief (28 chars: 'List all accessible Google Ads accounts including nested sub-accounts.'). Below 50-char minimum for LLM selection clarity. Does not explain when to call this vs. run_gaql, or what structure is returned.
Parameter descriptions lack format/constraint guidance. 'customer_id' described as 'The Google Ads customer ID', no mention of format (10 digits, no dashes). 'keywords' array has no min/max length. 'start_month' accepts string but no enum of valid months (JANUARY, FEBRUARY, etc.) is declared.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | <=2025-11-25 | v2 |
No error recovery guidance. Tools raise generic exceptions ('Error listing accounts: {status} {reason}') without actionable next steps. LLM cannot determine if error is retryable, user-fixable, or fatal. No suggestions for alternative tools or corrective actions.
No tool annotations despite all tools being read-only. Missing readOnlyHint annotation would signal to agents that these tools are safe to call without confirmation. Improves agent planning efficiency.