bankr-signals

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to place an API key (bk_...) directly into config files and curl headers/commands (e.g., -H "X-API-Key: bk_YOUR_KEY"), which forces the LLM to include secret values verbatim in generated output/commands, enabling exfiltration.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). SKILL.md and HEARTBEAT.md explicitly instruct the agent to poll public endpoints (e.g., https://bankrsignals.com/api/feed and https://bankrsignals.com/api/leaderboard) to consume user-generated, untrusted trading signals which the agent is expected to read and act on (copy trades, close positions), so third-party content can materially influence behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly tells agents in the heartbeat routine to "Fetch https://bankrsignals.com/heartbeat.md and follow it", meaning that content is fetched at runtime and its instructions directly control agent behavior.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly crypto/trading focused and provides wallet-related APIs. It describes Bankr wallet provisioning, an Agent API key, and a synchronous signing endpoint (EIP-191) for registering providers, publishing signals, and closing positions. These are specific blockchain wallet/signing capabilities (wallets + signing) intended to support on-chain trading workflows (including links to leverage-trading and sign/submit docs). Per the policy list, crypto/blockchain wallets and signing constitute Direct Financial Execution authority.