Single tool 'execute_skill_action' has critical definition gaps. Description is vague (79 chars, below 100-char minimum for clarity) and does not explain WHAT the tool does, WHEN to use it, or WHAT it returns. Input schema exists but parameter 'query_payload' lacks actionable description, it is a free-form string with no format, length, or content constraints. No output schema is documented; the response is a generic JSON object with 'status' and 'message' fields, but the LLM cannot predict structure or downstream chaining. No error handling guidance. Tool name 'execute_skill_action' is generic and does not follow verb_noun pattern (e.g., 'verify_hallucination', 'check_factuality'). The server claims to be a 'Multimodal Hallucination Factuality Verifier' but the tool does not reflect this domain, it is a catch-all governance operation. No evidence of idempotency, permission gates, or audit trails.

• Rename tool to 'verify_hallucination_factuality' or 'check_factuality_consistency' to match domain and follow verb_noun pattern. This enables LLMs to infer purpose from the name alone.
• Expand tool description to 100-200 chars. Example: 'Verify factuality of AI-generated statements against ground truth. Detects hallucinations, contradictions, and factual errors. Returns factuality score (0-1), detected hallucinations, and audit report URL. Use when validating agent outputs for accuracy.'
• Replace 'query_payload' with explicit parameters: 'ground_truth_context' (string, 10-5000 chars, required) and 'agent_statement' (string, 10-5000 chars, required). Add descriptions explaining what each field should contain.
• Document output schema with typed fields: 'verification_id' (string), 'factuality_score' (number 0-1), 'hallucination_detected' (boolean), 'contradiction_segments' (array of objects with 'claim', 'ground_truth', 'severity'), 'audit_report_url' (string, URI format).
• Add error handling: return structured errors with actionable guidance. Example: 'Invalid input: ground_truth_context must be 10-5000 characters. Received: 5 chars. Please provide more context.' Include retryability classification (retryable vs user-fixable).