The Agent Skills Directory

[SAFE]: The skill acts as a security and quality assurance tool using adversarial sub-agents to analyze code for vulnerabilities and errors.
[SAFE]: An indirect prompt injection surface is present as the skill processes target files provided by users. Evidence: Ingestion points occur in the Judge and Fix Agent templates via target description placeholders. Boundary markers include structured markdown headers to scope the target content. Capabilities are restricted to agent delegation for analysis and file modification. No specific sanitization of the target code is implemented, which is typical for development-focused orchestration skills.
[SAFE]: No instances of data exfiltration, credential exposure, or remote code execution were identified. All operations are confined to the platform's standard delegation tools.

judgment-day