youtube-video

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly requires a FREVANA_TOKEN to be provided via an environment variable or as a --token command-line argument, which instructs the agent to accept and pass a secret value (potentially embedding it verbatim in commands/requests), creating an exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.75). The required workflow calls an external Frevana endpoint and then reads the returned HTTP response body (outsider-authored content from a third-party service) as text via scripts/get_youtube_video.sh and prints it to stdout, which would be ingested into the agent’s LLM context if the agent captures stdout.