Paid Feedback Task for AI Agent / MCP Tool Users

We are looking for people who already use AI agent tools such as Claude Code, Cursor, n8n, Zapier MCP, MCP servers, or similar automation tools. We are building a control layer that sits between AI agents and MCP servers. McpSecRouter lets users decide which tools/actions are exposed to the agent before the model can see or use them. This is a paid product feedback / early design partner task. We want to understand whether this solves a real problem for people building or using AI agents, or whether it is just a “nice security feature” that is not urgent. Your task Briefly explain your current AI agent / MCP / automation workflow. Review or test MSR using a demo account. Record a short 5–10 minute screen recording while testing. Share honest feedback on what was useful, confusing, or missing. We are especially interested in practical feedback: when you would actually use this, which tools/actions you would want to control, and what would need to be added or improved before you would seriously use it. This is not a full development job. We are looking for practical feedback from real users of AI agent tools. Good fit You have used Claude Code, Cursor, MCP servers, n8n, Zapier, AI automation tools, or agent workflows. You have connected agents or automations to real tools/APIs. You can clearly explain your workflow. You are comfortable giving honest product feedback. Deliverables 5–10 minute screen recording Short written feedback Optional paid follow-up call if your feedback is especially relevant First task: fixed price (We are planning to hire 3–5 people for the first round of feedback.) Second task: This starts with a small paid feedback task. Strong participants may be invited to a paid follow-up video call and possible longer-term design partner work.