Coordinate three independent agents to validate outputs and issue blocking verdicts.
npx skills add https://github.com/mgsystemsdev/POD --skill review-mgsystemsdev
Please help me install this Agent Skill. Skill: review Source: https://github.com/mgsystemsdev/POD/tree/main/agent-system-base/.claude/skills/review Command: npx skills add https://github.com/mgsystemsdev/POD --skill review-mgsystemsdev
Gate every output through three independent hardening agents before anything ships to ensure safety, security, and compliance.
Run the tri-agent hardening review on a proposed change and observe the aggregated outcome before publishing.
High-intent search queries and answers about installing and using this skill.
A multi-agent security audit works by running three independent hardening agents—a devil-advocate, security-reviewer, and drift-detector—in parallel to validate proposed code changes, design proposals, or policy updates. They aggregate their findings into a single SHIP, REVISE, or BLOCK verdict to enforce compliance and safety.
A three-agent hardening review is a gating mechanism for software delivery where a devil-advocate, security-reviewer, and drift-detector concurrently audit an output. They consolidate their independent verdicts to determine if a proposed change, solution, or decision is safe to ship or requires revision.
Yes, you can use this hardening review process for design proposals and policy updates. The concurrent triage of the devil-advocate, security-reviewer, and drift-detector agents applies to any software delivery workflow or decision where risk controls, compliance, and safety matter.
The three independent agents aggregate their audit findings through a dedicated consolidation step that reviews the parallel verdicts from the devil-advocate, security-reviewer, and drift-detector. This aggregation produces a single SHIP, REVISE, or BLOCK decision, generating fix tasks automatically if a blocking verdict is necessary.
When a blocking verdict is triggered during a security review, the aggregation step generates specific fix tasks to address the risks identified by the hardening agents. This prevents the proposed code change, solution, or decision from shipping until the compliance and safety issues are resolved.
snake-dev-teamSearch codebases with parallel agents to locate files and gather context.
heyJordanParkerApply a 5-step framework to simplify software planning and reviews.
murphytruemanGenerate CI/CD pipeline configurations for design system quality checks.
Ed1s0nZAnalyze Android and iOS applications for security vulnerabilities using static and dynamic techniques.
