bankr-shopify

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime curl calls to Bankr's agent endpoints (e.g., https://api.bankr.bot/agent/prompt and https://api.bankr.bot/agent/job/{id}), which submit prompts and cause remote agent actions/settlement and are required for the Bankr bridge flows.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly includes Bankr bridge integrations that submit on-chain payment/settlement jobs. It documents calling Bankr's agent endpoints (POST https://api.bankr.bot/agent/prompt, GET /agent/job/{id}) and shows example prompts that instruct Bankr to "send 10 USDC to ${HANDLE} on base", "settle in USDC on base", and "send 100 LOYALTY to ${HANDLE} on base". It also lists MCP tools (bankr_agent_submit_prompt, bankr_agent_get_job_status, bankr_agent_cancel_job). Those are concrete crypto/blockchain payment operations (token transfers / settlement), not generic API callers or browser automation, so this skill provides direct financial execution capability.