An MCP Approval Server for AI Agents

Description and use

Impri’s approval server provides a safe mechanism for AI agents to propose actions, wait for human approval, and execute only after approval, ensuring auditability.

Problem

AI agents can perform destructive actions without a reliable human approval mechanism, leading to potential mistakes or malicious behavior.

Outcome

Agents can safely propose destructive actions, receive human approval, and execute only upon approval, ensuring auditability and preventing accidental or malicious changes.

Workflow

• Configure Impri MCP server in client config (set IMPRI_BASE_URL and IMPRI_API_KEY).
• Add Impri tools (impri_push_action, impri_await_decision, impri_report_result) to agent’s tool list.
• Wrap destructive actions (e.g., GitHub issue close) in a function that: 1. Calls impri_push_action to create pending action and notify reviewer. 2. Calls impri_await_decision to block until approved, rejected, or timeout. 3. If approved, perform the actual destructive action. 4. Call impri_report_result to log success or failure. 5. Return status to caller.

Evidence

Impri's MCP approval server (npx @impri/mcp) gives Claude Code and other MCP clients a real tool call to pause an agent and wait on a human's decision.

Source

impri_push_action, impri_await_decision, and impri_report_result show up as ordinary tools in the client's tool list.

Source

If the reviewer rejects it, or 30 minutes pass without a decision, closeIssueWithApproval returns without ever calling github.issues.update.

Source