visual-edit

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent/user to capture a server-minted bridgeToken and then embed that exact token value into shell commands / env var assignments (e.g., AGENT_NATIVE_BRIDGE_TOKEN="<bridgeToken from step 2>"), which requires the model to handle and output secret values verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). The visual-edit skill’s required runtime workflow reads a localhost app’s route/content by connecting to devServerUrl (e.g., /manifest.json) and rendering URL-backed iframe screens, which can include outsider-authored free text if the running app surfaces untrusted user input.