inbox-open-notification

Pass

Audited by Gen Agent Trust Hub on Jul 10, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • Dynamic Command Construction: The skill instructs the agent to execute shell commands such as open, xdg-open, or the GitHub CLI (gh) using a {url} placeholder. While this is the intended method for opening specific links, it represents a pattern where user-provided input is interpolated into a command line. The agent should be configured to validate that the input provided for {url} is a legitimate URL.
  • Standard Tool Utilization: The skill relies on established system utilities and the official GitHub CLI. These are well-known tools used for browser interactions and repository management in development environments.
  • Indirect Prompt Injection Surface: As the skill processes external data (GitHub URLs) and uses them in command-line arguments, there is a technical surface for indirect prompt injection.
  • Ingestion points: External URLs processed from GitHub notifications (referenced in SKILL.md).
  • Boundary markers: None explicitly defined for the {url} variable.
  • Capability inventory: Uses the execute tool to run shell commands.
  • Sanitization: No specific sanitization logic is described within the instruction file itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 10, 2026, 02:51 PM
Security Audit — agent-trust-hub — inbox-open-notification