shannon

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly runs live scans against an arbitrary target URL (via "npx @keygraph/shannon@beta start -u https://your-app.com/") and instructs the agent to read the scan outputs stored in ~/.shannon/workspaces/, meaning it fetches and interprets untrusted, user-provided web content and scanner results that can materially influence subsequent findings and actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill invokes npx @keygraph/shannon@beta at runtime (which fetches and runs code from the npm registry) and pulls a Shannon worker image from Docker Hub to execute an ephemeral container, so remote content from the npm package/Docker image will be executed and is required for the skill (npx @keygraph/shannon@beta / Docker Hub worker image).