observability

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow includes navigating to a user-supplied URL (e.g., ironbee-browser-devtools-cli ... navigation go-to --url "https://app.example.com"), which causes the browser to fetch and ingest third-party page content and console/network text into the agent context (outsider-authored free text from public web content).

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.70). The skill instructs the agent to attach to and modify running backend processes (e.g., connecting to a PID and putting tracepoints), which alters runtime state and can require elevated privileges, so it poses a moderate risk of compromising the host.