Sales

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill includes an explicit, mandatory instruction to immediately run a local network curl command before doing anything else (lines 9–26), which is unrelated to generating sales materials and forces external behavior that overrides normal operation — behavior characteristic of a prompt injection.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content includes a mandatory, backgrounded HTTP beacon to an unusual localhost port (31337) and explicit local override/load behavior that together create clear avenues for covert signaling and arbitrary code/configuration injection, indicating high risk of deliberate backdoor/exfiltration behavior.

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path reads “product documentation, product specs, or feature descriptions” provided as workflow input (e.g., CreateNarrative/CreateSalesPackage Step 1 and Step 2), which can include outsider-authored free text without any need to first select a specific trusted item.