ai-consultation

Code review with multiple AI perspectives validates decisions before commit. This Skill consults Codex, Gemini, and other models to surface security, performance, and architectural concerns automatically, reducing blind spots in peer review.

Yes. Architecture discussions benefit from structured multi-model evaluation to stress-test design choices. This Skill applies templated prompts and decision-logic guidance to architecture tasks, prioritizing findings across independent AI perspectives.

Debugging with AI consultation applies structured workflows to isolate root causes. This Skill coordinates Codex and Gemini evaluation with sandbox safety constraints, letting you request specific analysis—security audits, performance profiling—with tailored output formats.

Security audits run templated prompts across Codex and Gemini to catch authentication, data-handling, and injection vulnerabilities. This Skill sandboxes recommendations as read-only by default, requiring explicit approval before any write operations.

Yes. Refactoring validation uses multi-model consultation to check logic correctness and maintainability impact. This Skill provides decision-logic guidance and structured output to compare trade-offs across AI perspectives before committing changes.

Multi-model consultation requires explicit intent matching; the Skill validates that your request fits code review, security, or architecture scopes. Sandbox mode is read-only by default—write operations need explicit approval to prevent unintended modifications.