netbox-discovery

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime config manager can use config_manager.active: git to poll a Git repo and load selector.yaml/policy files; those files are outsider-authored free text that the agent ingests into its execution context (and thus into LLM-relevant processing) via the Git-based policy retrieval path.

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

• Potentially malicious external URL detected (high risk: 0.90). The agent's git config_manager polls remote Git repos at runtime (examples: https://github.com/org/policies.git and https://github.com/org/orb-policies.git), and the fetched selector.yaml and policy files directly control the agent's behavior/configuration, meeting the criteria for a high-confidence runtime external dependency.

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 repeatedly recommends running the agent with elevated privileges (e.g., --net=host, --privileged, -u root, sudo podman, CAP_NET_RAW) and advises bypassing rootless restrictions, which encourages actions that can bypass host security and compromise the machine state.