haproxy

Pass

Audited by Gen Agent Trust Hub on Mar 15, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides standard HAProxy management commands such as sudo systemctl reload haproxy and sudo socat. These are intended behaviors required for service administration and runtime interaction.
  • [CREDENTIALS_UNSAFE]: Includes a boilerplate admin:password string in the HAProxy stats dashboard configuration. This is a generic documentation placeholder and does not represent a sensitive secret leak.
  • [EXTERNAL_DOWNLOADS]: References official Docker images for haproxy:2.8 and nginx:alpine from well-known public registries.
  • [PROMPT_INJECTION]: The skill provides templates that incorporate variables, creating a surface for indirect injection if an agent interpolates unsanitized user data.
  • Ingestion points: Variable fields for IP addresses, domains, and paths in configuration snippets within SKILL.md.
  • Boundary markers: None present in the configuration examples.
  • Capability inventory: Subprocess execution for service management and socket interaction commands.
  • Sanitization: Not applicable as the skill provides static templates for demonstration purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 15, 2026, 11:09 AM
Security Audit — agent-trust-hub — haproxy