skills/nvidia/skills/hsb-app/Gen Agent Trust Hub

hsb-app

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands on remote targets via SSH using heredoc patterns to maintain state across sessions. It utilizes high-privilege operations, including 'sudo' for setup tasks and privileged Docker containers ('--privileged', '--net host') to access FPGA and sensor hardware. All major phases of operation require explicit user confirmation before proceeding.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by reading and processing external data from the remote repository and container logs to filter compatible applications and provide diagnostic guidance.
  • Ingestion points: Repository README files (e.g., examples//README), version files, and real-time Docker logs.
  • Boundary markers: No explicit delimiters or instructions are used to isolate untrusted remote content when interpolated into agent logic.
  • Capability inventory: The skill has access to the Bash tool for command execution and the Edit/Write tools for modifying code on the remote devkit.
  • Sanitization: There is no evidence of sanitization or filtering of content retrieved from the remote filesystem or container output before it is used for decision-making.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 07:42 PM
Security Audit — agent-trust-hub — hsb-app