pp-costco

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill tells the agent to capture the browser's short-lived idToken (localStorage.idToken/clientID) and "set it with auth set-token", which requires accepting and embedding the token verbatim in a command or config—exposing secrets through the agent's input/output.

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

• Third-party content exposure detected (medium risk: 0.65). The required workflow runs the costco-pp-cli commands (e.g., receipts, sync, raw) which fetch Costco account/receipt data at runtime and then returns that fetched text/JSON to the agent/LLM context via the CLI’s --agent/stdout output; this is outsider-authored content from Costco’s systems (not authored by the operating user).