owlp-cli

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill allows agent-driven CLI invocations that can require embedding user secrets verbatim (e.g., --auth-code, --mnemonic, wallet export-key) and explicitly contemplates the agent running those commands, which creates a high risk of secret exposure via the LLM output/context.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly a crypto wallet CLI with built-in transaction capabilities. It defines commands to create/import/export wallets and private keys and to send tokens (including a preview step and an execute step: owlp send ... --confirm). The command reference and workflows explicitly describe constructing, previewing, submitting, and polling blockchain transfers (NDJSON event stream for send, transaction history, tx detail). These are specific, purpose-built crypto/transaction operations (wallet management, signing/submitting transfers, off-ramp KYC for fiat withdrawals), so it grants direct financial execution authority.