universal-skill-architect

Warn

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: MEDIUMPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it is instructed to research 'emerging technologies' from external sources and use that information to draft new 'high-fidelity' skill instructions. An attacker could influence these external sources to inject malicious commands or logic into the generated skills, which the agent then automatically commits to its ecosystem.
  • Ingestion points: External research on 'emerging technologies' performed every 24 hours (SKILL.md).
  • Boundary markers: Absent; there are no instructions to sanitize or isolate external research data from the generation logic.
  • Capability inventory: Capability to automatically commit new skills to 11+ repositories and use spawn_agent (SKILL.md).
  • Sanitization: Absent; the skill does not include validation steps for the content harvested during research.
  • [COMMAND_EXECUTION]: The skill instructions mandate the autonomous creation and committing of code (SKILL.md files) to production repositories 'without user input'. This creates a risk of persistent, automated modifications to the agent's behavior and capability surface based on dynamically generated content.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 16, 2026, 03:06 AM
Security Audit — agent-trust-hub — universal-skill-architect