What Gets Installed

CodexFlow installs a repo-native workflow OS for OpenAI Codex.

CodexFlow is not a prompt pack. It installs structured, human-readable workflow files into your repository so Codex can start with context, follow project rules, use repeatable procedures, and leave behind cleaner handoffs.

CodexFlow is independent and is not affiliated with, endorsed by, or sponsored by OpenAI.

Installation model

CodexFlow has one paid install path in the current public offer:

CodexFlow is sold as a paid digital workflow product. Public demo materials show the install shape, while paid plans provide the signed activation and signed paid Individual payload.

AGENTS.md guidance

AGENTS.md gives Codex repo-level instructions.

It helps define:

• What the project is
• How the repo is structured
• Which conventions matter
• Which files or areas require extra care
• How Codex should plan before editing
• How testing and validation should be handled
• How handoffs should be written
• What “done” means for meaningful work

This is the foundation of the CodexFlow workflow OS.

Repo memory

Repo memory keeps durable context in the repository instead of leaving it buried in chat history.

CodexFlow memory can include:

• Project context
• Architectural decisions
• Known issues
• Prompt logs
• Work handoffs
• Risk notes
• Release notes
• Operating assumptions

The practical goal is to make future Codex sessions warmer, safer, and less repetitive.

Skills

CodexFlow skills are repeatable specialist workflows for common development tasks.

They are designed to help Codex work through areas such as:

• Performance
• SEO
• Security
• Supabase
• Vercel
• Deployment debugging
• React component refactoring
• Unit testing strategy
• Payment safety
• Authentication review
• Release readiness

Skills give Codex more than a general instruction. They give it task-specific operating procedures.

Slash-command workflows

Slash-command workflows turn repeated development rituals into reusable commands.

They help standardize:

• Planning
• Implementation
• Review
• Refactoring
• QA
• Release preparation
• Handoff
• Troubleshooting

Instead of rewriting the same prompt every time, you can reuse a structured workflow that lives with the repo.

QA guardrails

QA guardrails help keep AI-assisted work accountable.

CodexFlow can add structure for:

• Dry-runs
• Backups
• Risky-file detection
• Safe write plans
• Validation checklists
• Test planning
• Manual review boundaries
• Release checks

The goal is not to make AI edits magically safe. The goal is to make risks visible before they become expensive.

Shipping runbooks

Runbooks give Codex and humans known-good procedures for common operational work.

Examples include:

• CI repair
• Incident debugging
• Deployment troubleshooting
• Feature delivery
• Release preparation
• Repo audits
• Payment or authentication review
• Production-readiness checks

Runbooks reduce improvisation when the task has real consequences.

Handoff templates

Handoffs make each session easier to continue later.

A strong handoff captures:

• What changed
• Why it changed
• What was tested
• What was not tested
• What remains risky
• What should happen next
• Which files matter
• Which assumptions future sessions should preserve

CodexFlow treats handoff as part of the workflow, not an afterthought.

Public demo contents

The public demo provides read-only sample material so buyers can inspect the workflow shape before purchase.

Verified package-boundary checks include:

• CLI entry files required for paid setup
• License and notices
• No paid payload archives in static website paths
• No activation files in the public package
• No production signing keys in release artifacts

The release candidate package check verified a clean package boundary with no paid payload leakage.

Paid Individual payload contents

Personal and Standard customers receive the same paid Individual payload, governed by different license rights.

The paid Individual payload includes:

• Signed activation file
• Signed payload manifest
• Verified payload archive
• Payload hashes
• Paid agents
• Paid commands
• Paid skills
• Customer install documentation
• Premium workflow files emitted with license-aware watermark metadata

The current paid Individual payload release contains:

• 8 paid agents
• 10 paid commands
• 20 paid skills
• 38 paid workflow asset files in the signed payload archive

Watermarked premium output

Paid premium output includes license metadata so usage rights are visible without exposing raw secrets.

Watermark metadata can include:

• CLI version
• Generated timestamp
• License tier
• Commercial-use status
• Truncated license identifier hash
• License-key hash prefix
• Payload identifier and version

Watermarks do not include raw license keys, raw customer emails, raw license IDs, private keys, or customer source-code excerpts.

Local-first by design

CodexFlow installs workflow files into your repository. The files are designed to be:

• Human-readable
• Git-reviewable
• Versionable
• Editable by the repository owner
• Visible in diffs
• Easy to remove or revise

The CodexFlow CLI does not transmit usage telemetry in this release. The commerce website records operational data needed for payment, fulfillment, access, fraud prevention, support, and dispute evidence.

What CodexFlow does not install

CodexFlow does not install:

• A replacement for Codex
• A hosted coding agent
• A secret background service
• A production monitoring agent
• A team permissions system
• A guarantee that AI-generated code is correct
• Unlimited commercial rights under the Personal tier

CodexFlow gives your repo a workflow system. You still review, test, and own the code you ship.

Review before committing

After installation or payload emission, always review the Git diff:

git status
git diff

Commit only the workflow files you want to keep.

Start after purchase

Get CodexFlow
Compare plans