Prompt snapshot reader
Codex CLI system prompt
This reader maps a small, source-verbatim excerpt from an official Codex repository prompt file and keeps product, model, surface, and runtime scope separate.
Publication context
Read the provenance first
Apache-2.0 repository source supports limited attributed republication; the page does not claim to represent a complete runtime assembly.
- Source tier
- Tier B · first-party open source
- Confidence
- high
- Reviewed
- Aug 10, 2026
- Completeness
- partial
- Checksum
- sha256:0911121f55c4f…
- Raw artifact
- Not publicly reproduced
Three source-verbatim excerpts are organized under stable anchors. Search, copying, and wrapping enhance the static document.
Identity and scope
You are GPT-5.2 running in the Codex CLI, a terminal-based coding assistant. Codex CLI is an open source project led by OpenAI. You are expected to be precise, safe, and helpful.
Reading note. The excerpt identifies the surface as Codex CLI and separates it from the older Codex language-model name. It does not establish the complete instruction stack.
Harness context and communication
Your capabilities: - Receive user prompts and other context provided by the harness, such as files in the workspace. - Communicate with the user by streaming thinking & responses, and by making & updating plans.
Reading note. The source describes context supplied by the harness and names planning as part of the assistant's working interface. It does not document all possible context sources.
Tool use and approval boundary
- Emit function calls to run terminal commands and apply patches. Depending on how this specific run is configured, you can request that these function calls be escalated to the user for approval before running.
Reading note. The excerpt establishes a configuration-dependent tool and approval boundary. It should not be read as a guarantee that every run exposes the same tools or permissions.
Evidence and provenance
Tier B · first-party open source: OpenAI Codex prompt file — First-party open-source source; Promptcred republishes a limited excerpt with attribution.
Snapshot evidence
What this snapshot includes and leaves out.
- Product surface
- Codex CLI
- Model context
- GPT-5.2This model name comes from the selected public prompt file and is not generalized to every Codex runtime.
- Checksum scope
- Normalized three-section public excerpt package, UTF-8, LF line endings.
- Raw mode
- The public package retains a normalized limited excerpt only; it does not bundle the complete upstream file.