awesome-codex-subagents

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download configuration files from the vendor's GitHub repository (github.com/VoltAgent/awesome-codex-subagents.git) via git clone.
  • [COMMAND_EXECUTION]: Includes standard filesystem commands (mkdir, cp, mv) for installing and managing subagent configurations in the user's local directory.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: Promotes secure development practices by explicitly instructing users to use environment variables for sensitive data (API keys, database URLs) and warning against hardcoding secrets.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes an architecture where subagents process project data (code reviews, dependency audits). While this creates a theoretical surface for indirect prompt injection from malicious project files, the skill provides specific instructions to mitigate risk, such as restricting sandbox modes (read-only) and using environment variables for secrets.
  • [PRIVILEGE_ESCALATION]: Mentions various sandbox levels for subagents (read-only, workspace-write, full), advising caution when using elevated permissions (full).
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 04:50 PM