MCP server for Tuteliq — 50 AI-powered tools for child safety, fraud detection, synthetic content forensics, identity verification, grooming, bullying, sextortion, document analysis, and content moderation. Interactive UI widgets for Claude, Cursor, and MCP-compatible AI assistants.
Tuteliq MCP has 14 tools with solid naming conventions (all verb-prefixed: list_, create_, update_, delete_, test_, get_, analyze_, generate_). Descriptions are present for all tools and range from adequate to comprehensive. However, there are significant gaps in input schema completeness, parameter descriptions, and output schema documentation. All tools have descriptions ≥20 characters, which prevents hard cap violations. The analyze_emotions and get_action_plan tools have the most detailed parameter descriptions and complex schemas. Several tools (especially admin tools like test_webhook, regenerate_webhook_secret) lack detailed parameter descriptions. No explicit output schemas are documented in the provided source code. Error handling guidance is not evident from tool definitions.
Analyze emotional content and mental state indicators. Identifies dominant emotions, trends, and provides follow-up recommendations. Pass either `content` (single text) or `messages` (a conversation with senders) — messages enable per-speaker trend analysis.
Create a new webhook endpoint.
Permanently delete a webhook.
Generate a comprehensive incident report from a conversation.
Generate age-appropriate guidance and action steps for handling a safety situation.
Get available pricing plans for Tuteliq.
No documented output schemas for any tools. LLMs cannot plan downstream calls or extract return values without knowing what fields to expect.
Admin tools (test_webhook, regenerate_webhook_secret, get_pricing, get_pricing_monthly) have minimal descriptions or parameter details. Descriptions under 70 chars lack context for LLM tool selection.
No explicit error handling guidance in tool definitions. Tools do not indicate what errors are retryable, user-fixable, or fatal, or what recovery steps an agent should take.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | B | 75 | 2025-06-18+ | v2 |
| 2026-03-09 | B | 71 | - | v1 |
Get detailed pricing plans with features, limits, and rate caps.
Get usage broken down by tool/endpoint for a specific date.
Get daily usage history for the past N days.
Monthly Usage
List all webhooks configured for your account.
Regenerate a webhook signing secret. Invalidates the previous secret immediately.
Send a test payload to a webhook to verify it is working correctly.
Update an existing webhook configuration.
Destructive tools (delete_webhook, regenerate_webhook_secret) lack confirmation or dry-run support. Agents could accidentally delete webhooks or invalidate secrets without safeguards.
Parameter constraints not fully specified. For example, 'days' parameter in get_usage_history states '1-30, default: 7' in description but no JSON Schema min/max constraints visible. 'date' parameter in get_usage_by_tool states 'YYYY-MM-DD format' in description but no pattern constraint visible.
analyze_emotions tool uses 'context' as an object without specifying required or expected fields. Tool accepts either 'content' or 'messages' but mutual exclusivity not enforced or documented in schema.