ptz-onvif

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt repeatedly shows and instructs embedding plaintext credentials (e.g., --password 'S3cret' and rtsp://admin:S3cret@...) in commands and URLs, which requires the LLM to output secret values verbatim and is therefore insecure.

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime LLM context exposure is via outsider-authored free text returned by the camera over the network: e.g., discover/info/streams/snapshot call soap_request() and then parse/print XML fields like XAddrs, Manufacturer, Model, Uri, etc., which are attacker-controlled by the remote camera and become readable text in the agent’s output/LLM context.