backchannel_start

THE RISK

High Risk

Why backchannel_start needs a policy

Although the tool itself doesn't complete a financial transaction, it initiates external authentication mechanisms (push notifications, SMS) that represent triggered operations whose effects depend on arguments. This exceeds passive data retrieval (Read) and non-destructive modification (Write).

From the tool's definition Tool description states it 'Sends a push to Nubank app (CIBA) or triggers an OTP SMS' — these are external operations that initiate out-of-band authentication flows with side effects outside the system.

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

POLICY

How to control backchannel_start

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

policy.json

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

backchannel_start 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 Mcp Afip — 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 Mcp Afip tools

Financial authorize_transaction Authorize a transaction (reserve funds without capturing) via Braintree GraphQL authorizeP Financial cancel_collection Cancel a scheduled collection request Financial cancel_credit Cancel an authorized-but-uncaptured credit payment. Financial cancel_payment Attempt to cancel a payment that has not yet settled. Iniciador endpoint: POST /payments/{ Financial cancel_payment_intent Cancel a payment intent that has not yet been captured. Fails on already-captured intents; Financial cancel_recurrence Cancel an active recurrence. Future charges will stop; already-charged cycles are untouche Financial cancel_subscription Cancel a subscription by ID Financial revoke_token Revoke an issued access_token or refresh_token at POST /v1/token/revoke. Use to terminate

All 1300 Mcp Afip 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 Mcp Afip → 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 backchannel_start

What does the backchannel_start tool do? +

Start a CIBA / OTP pre-authorization for a shopper. Sends a push to Nubank app (CIBA) or triggers an OTP SMS. Returns an auth_req_id/ticket to complete later. It is categorised as a Execute tool in the Mcp Afip MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on backchannel_start? +

Register the Mcp Afip MCP server in PolicyLayer and add a rule for backchannel_start: 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 Mcp Afip. Nothing to install.

What risk level is backchannel_start? +

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

Can I rate-limit backchannel_start? +

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

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

backchannel_start is provided by the Mcp Afip MCP server (codespar/mcp-dev-latam). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Enforce policy on every Mcp Afip tool call.

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

CHECK YOUR STACK →