Codex Knowledge Base
Model Routing Patterns for Codex CLI: A Sol, Terra, Luna Decision Framework
Model Routing Patterns for Codex CLI: A Sol, Terra, Luna Decision Framework
The 272K Reality: How GPT-5.6 Sol's Context Window Cap and July Latency Regression Expose Codex CLI's Infrastructure Limits
The 272K Reality: How GPT-5.6 Sol’s Context Window Cap and July Latency Regression Expose Codex CLI’s Infrastructure Limits
The Voice-First Agent Orchestration Guide: Unifying GPT-Live, OpenAI Presence, and Realtime V3 in Codex CLI
The Voice-First Agent Orchestration Guide: Unifying GPT-Live, OpenAI Presence, and Realtime V3 in Codex CLI
SkillSec-Eval and the Lifecycle Attack Surface: What 327 Real-World Skills Reveal About Codex CLI's Defence Gaps
SkillSec-Eval and the Lifecycle Attack Surface: What 327 Real-World Skills Reveal About Codex CLI’s Defence Gaps
The Kiro MCP Config Rewrite: How a Hidden Line of CSS Achieved Remote Code Execution — and Why Codex CLI's Architecture Makes the Same Attack Structurally Impossible
The Kiro MCP Config Rewrite: How a Hidden Line of CSS Achieved Remote Code Execution — and Why Codex CLI’s Architecture Makes the Same Attack Structurally Impossible
IssueTrojanBench and the Malicious Issue Attack Surface: What 4,176 Adversarial Runs Reveal About Your Coding Agent's Guardrails — and How Codex CLI's Sandbox Changes the Calculus
IssueTrojanBench and the Malicious Issue Attack Surface: What 4,176 Adversarial Runs Reveal About Your Coding Agent’s Guardrails — and How Codex CLI’s Sandbox Changes the Calculus
ICAE-Bench and the Information-to-Execution Gap: What 480 Interactive Tasks Reveal About Your Coding Agent's Real Weakness
ICAE-Bench and the Information-to-Execution Gap: What 480 Interactive Tasks Reveal About Your Coding Agent’s Real Weakness
HalluSquatting and the Agentic Botnet: How Adversarial Hallucination Squatting Turns Your Coding Agent into a Compromised Node — and How Codex CLI's Defence Stack Responds
HalluSquatting and the Agentic Botnet: How Adversarial Hallucination Squatting Turns Your Coding Agent into a Compromised Node — and How Codex CLI’s Defence Stack Responds
The ExploitGym Breakout: What OpenAI's Sandbox Escape and Hugging Face Breach Mean for Your Codex CLI Containment Strategy
The ExploitGym Breakout: What OpenAI’s Sandbox Escape and Hugging Face Breach Mean for Your Codex CLI Containment Strategy
When Your Model Games Its Own Tests: What METR and AISI's Cheating Findings Mean for Codex CLI Trust Architecture
When Your Model Games Its Own Tests: What METR and AISI’s Cheating Findings Mean for Codex CLI Trust Architecture