Codex Process Jobs v0.2.1

Codex Process Jobs v0.2.1

This patch release improves update safety and makes installation provenance easier to inspect without exposing local paths.

Added

• codex-process-jobs doctor --provenance reports the release, runtime snapshot, validated cache generations, upstream repository, and editable-checkout status. It is read-only and path-redacted.

Changed

• Installation and update guidance now requires reviewing every CPJ hook definition and its referenced source through /hooks after every update.
• Users approve only definitions Codex marks new or changed; retained trust must still be verified.
• Release-version assertions now derive from package metadata, reducing mechanical release-only edits.
• The maintainer release runway now distinguishes merged source, immutable GitHub releases, Homebrew formula publication, per-host plugin updates, and client refresh.

Upgrade

After the Homebrew formula update is available:

brew update
brew upgrade codex-process-jobs
codex-process-jobs update --agent-policy none
codex-process-jobs update --apply --agent-policy none

Use the host's existing global or project --project-root /absolute/path policy instead of none where applicable. Review the preview before applying it. Then restart Codex clients, review /hooks, and start a fresh task.

Full changes: v0.2.0...v0.2.1

Codex Process Jobs v0.2.0

First versioned public release of Codex Process Jobs.

Highlights

• Durable detached execution for builds, tests, inference runs, repairs, and other long local processes
• Conversational completion relays for Codex App, Codex CLI, VS Code, and mobile-driven remote tasks
• Lightweight status, bounded tail/result inspection, and process-group cancellation
• Goal-aware hard turn boundaries that prevent replacement polling loops
• Explicit Bash, POSIX shell, and direct-argv execution modes
• Preview-first transactional installation with optional global, project, or no AGENTS.md policy
• Consent-gated hooks, strict persisted-state validation, and untrusted-output boundaries
• macOS and Linux support with no third-party runtime dependencies

Install with Homebrew

brew install joelfarthing/tap/codex-process-jobs
codex-process-jobs install

Review the read-only plugin installation preview, choose global, project, or none for the optional agent policy, and add --apply only after approving the plan.

See the README, changelog, and security model.

Community beta: this is an independent project, not an official OpenAI plugin.

Codex Process Jobs v0.1.0

Initial public beta of Codex Process Jobs: a dependency-free Codex plugin for running ordinary macOS and Linux commands as durable detached process jobs instead of holding an agent turn open.

Highlights

• Durable detached execution with process identity, bounded stdout/stderr, terminal status, and safe process-group cancellation.
• Namespaced start, status, tail, result, and cancel skills for CMake builds, long tests, inference runs, data processing, and repair utilities.
• Best-effort conversational completion across Codex App, CLI, VS Code, and remote/mobile-driven hosts, backed by durable status and result retrieval.
• Consent-gated PostToolUse, Stop, and UserPromptSubmit hooks; the installer never trusts hooks automatically.
• Critical-job safeguards, explicit Goal integration, completion batching, proactive bounded result inspection, and preservation of prior cache generations for open tasks.
• Two-phase installer with an explicit choice of global, project, or no AGENTS.md adoption policy.
• 148 tests plus a macOS/Ubuntu, Node.js 18/22 CI matrix.

Install with Codex

Tell Codex:

Install joelfarthing/codex-process-jobs from GitHub. Run the installer's read-only preview first and describe every local change. Then ask whether I want the optional AGENTS.md policy globally, in one project, or not at all. Do not apply the installation until I approve the preview and policy scope.

After installation, restart the Codex client and explicitly review and approve the plugin's hook hashes in /hooks.

Important limitations

• This is an independent community beta, not an official OpenAI plugin.
• Automatic conversational wake uses experimental local Codex transports and is best-effort; durable status and result retrieval remain authoritative.
• Commands receive no interactive stdin and must remain finite foreground processes. Servers, watchers, daemonized commands, and external fire-and-exit launchers need another lifecycle mechanism.
• Windows is not currently supported.
• CPJ primarily improves workflow quality by releasing the conversation. Token usage can be lower, neutral, or higher depending on the foreground baseline and result-inspection behavior; no universal savings percentage is claimed.

See the README, full changelog, and security model for details.