agent-creator
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as an interactive guide for generating application-specific configuration files (agent profiles). It follows a confirmation-based workflow where the user reviews the proposed configuration before it is written to disk.
- [SAFE]: Included reference materials provide clear guidance on security-enhancing features, such as setting sensitive tools (like bash) to 'ask' or 'never' approval states and implementing regex-based command filtering.
- [DATA_EXFILTRATION]: The skill manages local configuration storage within a specific application directory (~/.aider-desk/agents/). No patterns of credential harvesting or unauthorized remote data transmission were identified.
- [COMMAND_EXECUTION]: While the skill assists in configuring agents that may use shell tools, it does not perform arbitrary command execution itself. It explicitly documents methods to restrict shell access for the agents it helps create.
Audit Metadata