Codex Knowledge Base
PMCoder and the Plan-Memory Coupling: Why Bidirectional Phase-Aware Retrieval Beats Flat Context — and How to Wire It into Codex CLI
PMCoder couples hierarchical planning with episodic memory so that the current phase shapes what the agent remembers and accumulated evidence triggers replanning. Here is how to map that architecture to...
OneDayAgent and the Long-Horizon Harness: Why Decompose-Verify-Repair Beats Growing a Single Session — and How to Wire It into Codex CLI
OneDayAgent and the Long-Horizon Harness: Why Decompose-Verify-Repair Beats Growing a Single Session — and How to Wire It into Codex CLI
LongHorizon-Harness and the Manage-Execute-Audit Loop: Why Fresh-Context Execution with Verified State Beats Growing a Single Session — and How to Wire It into Codex CLI
LongHorizon-Harness and the Manage-Execute-Audit Loop: Why Fresh-Context Execution with Verified State Beats Growing a Single Session — and How to Wire It into Codex CLI
LivePlan and Online Corrective Steering: Why Deterministic Drift Detection Plus LLM Advisory Beats Global Replanning — and How to Wire It into Codex CLI
LivePlan and Online Corrective Steering: Why Deterministic Drift Detection Plus LLM Advisory Beats Global Replanning — and How to Wire It into Codex CLI
The IDK Filter and the Confident Hallucination Problem: Why Your Coding Agent Calls the Wrong Tool With Full Confidence — and How Uncertainty-Aware Filtering Maps to Codex CLI
The IDK Filter and the Confident Hallucination Problem: Why Your Coding Agent Calls the Wrong Tool With Full Confidence — and How Uncertainty-Aware Filtering Maps to Codex CLI
Don't Blame the Model: How Scaffolding Evolution Shapes Coding Agent Quality — and What It Means for Codex CLI Users
Don’t Blame the Model: How Scaffolding Evolution Shapes Coding Agent Quality — and What It Means for Codex CLI Users
Cross-Benchmark Generalization in Agentic RL: Why Training on Office Tasks Makes Coding Agents Better — and What It Means for Codex CLI Model Selection
Cross-Benchmark Generalization in Agentic RL: Why Training on Office Tasks Makes Coding Agents Better — and What It Means for Codex CLI Model Selection
Codex CLI v0.147.0: Portable Agent Plugins, Multi-Catalog Federation, and the --approve-for-me Flag
Codex CLI v0.147.0: Portable Agent Plugins, Multi-Catalog Federation, and the –approve-for-me Flag
Cheap Code, Costly Judgment: Governance Conversion for Agentic Software Engineering with Codex CLI
Cheap Code, Costly Judgment: Governance Conversion for Agentic Software Engineering with Codex CLI
Addressable Recall Compaction: Why Lossless Context Management Changes the Game for Codex CLI Long Sessions
Addressable Recall Compaction: Why Lossless Context Management Changes the Game for Codex CLI Long Sessions