claude_request_async

THE RISK

High Risk

Why claude_request_async needs a policy

This tool triggers execution of Claude Code CLI commands in a background context. While the tool itself does not directly perform destructive or financial actions, it invokes arbitrary Claude Code operations whose side effects depend entirely on what the user requests.

From the tool's definition Tool description states it 'Start[s] a Claude Code CLI request as a durable background job.' The ability to start requests to Claude Code CLI and trigger external operations whose effects depend on the request arguments qualifies as Execute.

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

POLICY

How to control claude_request_async

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

policy.json

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

claude_request_async 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 LLM CLI Gateway — 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 LLM CLI Gateway tools

Destructive llm_job_cancel Cancel a running gateway async or deferred-sync job by jobId. Destructive session_clear_all Delete all gateway session records, optionally scoped to one provider. Destructive session_delete Delete a gateway session record by ID (also removes any gateway-owned worktree attached to Execute claude_request Run a Claude Code CLI request synchronously (when async jobs are enabled, auto-defers to a Execute cli_upgrade Plan (dryRun, default true) or execute an upgrade for one provider CLI using its native up Execute codex_fork_session Fork an existing Codex session into a new branch (codex fork <ID|--last>) and run a prompt Execute codex_request Run an OpenAI Codex CLI request synchronously (when async jobs are enabled, auto-defers to Execute codex_request_async Start an OpenAI Codex CLI request as a durable background job. Poll with llm_job_status, c

All 46 LLM CLI Gateway 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 LLM CLI Gateway → 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 claude_request_async

What does the claude_request_async tool do? +

Start a Claude Code CLI request as a durable background job. Poll with llm_job_status, collect with llm_job_result. It is categorised as a Execute tool in the LLM CLI Gateway MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on claude_request_async? +

Register the LLM CLI Gateway MCP server in PolicyLayer and add a rule for claude_request_async: 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 LLM CLI Gateway. Nothing to install.

What risk level is claude_request_async? +

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

Can I rate-limit claude_request_async? +

Yes. Add a rate_limit block to the claude_request_async 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 claude_request_async completely? +

Set action: deny in the PolicyLayer policy for claude_request_async. 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 claude_request_async? +

claude_request_async is provided by the LLM CLI Gateway MCP server (llm-cli-gateway). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every LLM CLI Gateway tool call.

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

CHECK YOUR STACK →