cdf-event-video-generation

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of kubectl and curl to manage the lifecycle of Airflow DAGs and inspect Kubernetes cluster state. These operations are aligned with the skill's primary purpose of workflow orchestration and follow established management patterns.
  • [DYNAMIC_EXECUTION]: Several procedures use python3 -c or shell heredocs to execute inline Python snippets for parsing JSON output from kubectl and curl. This is used for diagnostic checks and data processing, such as verifying the presence of secrets without exposing their values.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external data through JSON configuration files (e.g., variable-distribution.json) and result manifests (e.g., dataset.json). This attack surface is mitigated by the included scripts/payload.py script, which performs strict schema validation and type checking on input data.
  • [EXTERNAL_DOWNLOADS]: The skill fetches configuration and documentation from NVIDIA's official GitHub repositories. These references target trusted vendor resources and provide necessary technical context for the agent's operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 03:52 PM