skills/sehoon787/my-claude/omc-teams/Gen Agent Trust Hub

omc-teams

Warn

Audited by Gen Agent Trust Hub on Mar 24, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Potential shell command injection via user-supplied task strings. The skill constructs shell commands such as omc team <N>:<agent> "<task>" by interpolating a user-provided task description. Without rigorous sanitization or escaping of shell metacharacters, an attacker could craft a task string that executes arbitrary commands on the host system.
  • [PROMPT_INJECTION]: Indirect prompt injection surface through task delegation. The skill functions as an orchestrator that passes untrusted task data to external AI CLI tools (Claude, Codex, Gemini). This creates a surface where adversarial instructions embedded in the task could manipulate the behavior of the spawned worker agents.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 24, 2026, 07:43 AM
Security Audit — agent-trust-hub — omc-teams