The Agent Skills Directory

[SAFE]: The skill contains purely instructional content regarding the deployment and configuration of the DOCA Argus Service container. No malicious patterns such as prompt injection, data exfiltration, or persistence mechanisms were found.
[NO_CODE]: The skill does not include any executable scripts (Python, Node.js, or Shell). It strictly provides documentation and guidance, as confirmed by the documentation_only: true flag in its metadata.
[INDIRECT_PROMPT_INJECTION]: Although the skill describes a service that monitors untrusted host data, it implements strong safety protocols. It mandates a calibration period for tuning detection policies and requires end-to-end smoke testing (Argus -> Forwarder -> SIEM) to establish a baseline before production use, minimizing the risk of automated obedience to embedded data instructions.
[EXTERNAL_DOWNLOADS]: The skill references pulling container images from NVIDIA NGC and forwarding findings to well-known SIEM services (Splunk, ELK, Microsoft Sentinel). These are recognized vendor resources and established services, which are treated as safe within their documented operational contexts.

doca-argus