cursor_recover_task

THE RISK

High Risk

Why cursor_recover_task needs a policy

This tool re-executes a previously failed task by relaunching it, which constitutes triggering an external operation. The 'relaunching' action is an Execute-category operation. Severity is medium because misuse could cause unintended task re-execution or resource consumption, but the blast radius is limited to the task/agent coordination scope.

From the tool's definition 'Recover a failed or stale delegated Cursor task by relaunching it with the latest shared context'

Documented attack patterns abuse exactly the kind of access cursor_recover_task gives an agent:

POLICY

How to control cursor_recover_task

PolicyLayer is an MCP gateway — it sits between your AI agents and Agent Orchestration, and nothing reaches the server without passing your rules. This is the rule we recommend for cursor_recover_task:

policy.json

{
  "version": "1",
  "default": "deny",
  "tools": {
    "cursor_recover_task": {
      "limits": [
        {
          "counter": "cursor_recover_task_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

cursor_recover_task stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.

Create a free account and register Agent Orchestration — nothing to install.
Add this policy — paste it, or build it visually.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

RATE-LIMIT THIS TOOL →

Free to start. No card required.

EXPLORE

Related tools and policies

More Agent Orchestration tools

Destructive memory_delete Delete a value from shared memory. Destructive agent_unregister Unregister this agent and release all held locks. Call at the end of your session. Execute bootstrap Initialize agent session: register (if needed), get current focus, pending tasks, and rece Execute lock_acquire Acquire a lock on a resource to prevent concurrent access. Write claim_todo FOR SUB-AGENTS: Register yourself AND claim a specific task in one call. Use this when you Write cursor_delegate_task Delegate a task to Cursor CLI and store session metadata for later resume/sync. Write cursor_sync_task Refresh a delegated Cursor task, sync the latest provider status, and persist knowledge ba Write research_ready Mark research as complete for a task. Validates that all required research items are docum

All 35 Agent Orchestration tools →

Execute tools on other servers

Android Forensics ADB MCP Server adb_shell_command EMBA-MCP run_emba_scan Frida Game Hacking MCP remove_breakpoint Procmon request_elevation

Go deeper

High-risk tools in Agent Orchestration → Every high-risk tool this server exposes, in one view.
All high-risk MCP tools → Every high-risk tool across the catalogue, and the policy pattern that contains them.
The MCP Attack Database → Documented attack patterns against MCP deployments — and the policies that stop them.
Agent sandbox → Bounding what an agent can execute and where the boundary should sit.
Least privilege for MCP → Granting agents the minimum tool surface the task needs.
Rate Limiting MCP Tool Calls: A Practical Guide →
MCP Security: Why Prompt Guardrails Aren't Enough →

FAQ

Questions about cursor_recover_task

What does the cursor_recover_task tool do? +

Recover a failed or stale delegated Cursor task by relaunching it with the latest shared context. It is categorised as a Execute tool in the Agent Orchestration MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on cursor_recover_task? +

Register the Agent Orchestration MCP server in PolicyLayer and add a rule for cursor_recover_task: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Agent Orchestration. Nothing to install.

What risk level is cursor_recover_task? +

cursor_recover_task is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit cursor_recover_task? +

Yes. Add a rate_limit block to the cursor_recover_task rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.

How do I block cursor_recover_task completely? +

Set action: deny in the PolicyLayer policy for cursor_recover_task. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.

What MCP server provides cursor_recover_task? +

cursor_recover_task is provided by the Agent Orchestration MCP server (madebyaris/agent-orchestration). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Agent Orchestration tool call.

Start from Agent Orchestration, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.

CHECK YOUR STACK →