Raw apply_patch — sends Codex's Lark grammar as an OpenAI custom tool, so patches are not JSON-wrapped.
Capability-based activation — requires openai-codex-responses or openai-responses plus model.compat.supportsOpenAIGrammarTools === true; model names alone are never enough.
Safe local mutation — patches are limited to 1 MiB, target files to 64 MiB, stay under Pi's current working directory, reject symlink escapes, use descriptor-anchored no-follow operations on Linux, fail closed elsewhere, preflight all hunks, and serialize writes with Pi's mutation queue.
Model switching — supported models replace Pi's edit and write tools with apply_patch; other active tools are preserved. Switching back restores only the file tools that were active before the switch.
Sequential patch calls — the extension marks patch execution sequential and disables provider-side parallel tool calls when the patch tool is active.
Streaming progress — while a patch is generated, the TUI shows a live, color-coded glimpse of the content being written (new-file content, or +/- lines for updates) plus a running +added -removed tally and a per-file roster for multi-file patches. It reuses Pi's shared diff rendering and mirrors the built-in write/edit previews; patch execution is unchanged.

Installation

pi install npm:pi-codex-tools

For a one-off run:

pi -e /path/to/pi-mono/packages/pi-codex-tools

Scope decisions

The current Codex source does not define separate read_file or write_file tools: file inspection is normally done through shell commands and file mutation through apply_patch. This package keeps Pi's bounded read and bash tools, and uses apply_patch in place of Pi's edit and write tools for supported models. Because Pi does not provide Codex's OS-level filesystem sandbox, apply_patch runs only on Linux and fails closed on unsupported platforms. It also requires a Pi model runtime that advertises compat.supportsOpenAIGrammarTools; older runtimes leave the tool inactive.

Codex surface	Decision
`apply_patch`	Included; it is a materially different freeform grammar tool.
`shell_command`	Deferred; Pi already has the bounded shell backend, while a faithful adapter needs Codex's approval and working-directory contract.
`exec_command` + `write_stdin`	Deferred; persistent PTY sessions need a separate process/session design.
`view_image`	Deferred; Pi's `read` already sends supported images as attachments.
`update_plan`	Deferred; it is workflow metadata rather than a capability-specific file tool.
Code Mode `exec` + `wait`	Deferred; it requires a real sandbox for model-authored JavaScript, not Node's ordinary `vm` wrapper.

These choices are based on the Codex tool specifications in codex-rs/core/src/tools, the model profiles in codex-rs/models-manager/models.json, and the Code Mode protocol. They intentionally keep this package focused on the one tool with a distinct transport and model-facing contract.

Compatibility notes

apply_patch is line-oriented rather than byte-oriented:

*** Add File requires at least one + line and writes a trailing newline. A +-only hunk creates a one-newline file, not a zero-byte file.
Updates produce a trailing newline for non-empty output, so updating a file that lacks one may add it.
Existing CRLF line endings are preserved when detected.
Use bash when exact byte-level output or a truly empty file is required.

These behaviors intentionally match Codex apply_patch.

The provider contract is runtime-specific: use Pi 0.83.0 or newer for OpenAI grammar-tool support. For a manual smoke test, start Pi with this extension and a model that advertises supportsOpenAIGrammarTools, then verify that a file change appears as an apply_patch call and not as edit, write, or bash.

Development

npm run -w packages/pi-codex-tools check
npm test -w packages/pi-codex-tools
npm run -w packages/pi-codex-tools pack:dry-run

Install/update telemetry is disabled in CI and can be disabled with PI_OFFLINE=1, PI_TELEMETRY=0 or PI_TELEMETRY=false, or Pi's enableInstallTelemetry: false setting. See SECURITY.md.