subagent-dev
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a methodical 'Subagent-Driven Development' process focusing on task isolation and multi-stage validation.
- [SAFE]: Explicit guardrails are provided within the instructions to enforce security best practices, including limiting file system access to necessary directories and prohibiting destructive commands like 'rm' without authorization.
- [SAFE]: Network access for subagents is restricted to specific, pre-defined scopes, reducing the risk of unauthorized data exfiltration.
- [SAFE]: No evidence of prompt injection, hidden obfuscation, or malicious remote code execution was found. The logic focuses on legitimate development automation and quality assurance.
Audit Metadata