zenigame-codex-review

Automate Codex review sessions by applying standardized frontmatter rules that enforce mission constraints and prevent prompt injection. This orchestrates codex-vscode calls, restricts file writes, and guides the Codex within predefined safeguards for consistent, safe analysis.

Prevent prompt injection during code analysis by enforcing a central mission with explicit restrictions. This framework requires explicit frontmatter info to constrain behavior, restricts unsafe file writes, and standardizes prompts to block injection attempts.

Audit Codex reasoning traces and prompts by preserving session state, decision logs, and analysis prompts. This skill enables governance by maintaining an auditable record of reasoning traces and frontmatter-driven constraints across iterative review modes.

Yes, you can use codex-vscode for iterative session-based code reviews. This framework supports both one-shot and session-based review modes, allowing iterative refinement while maintaining safety rules, mission constraints, and auditability throughout the session.

Codex analysis output suffers from inconsistent prompt framing due to unmanaged prompt usage. Applying a standardized orchestration framework with explicit frontmatter-driven rules ensures consistent framing, enforces mission constraints, and eliminates unsafe prompt variations.