-
Notifications
You must be signed in to change notification settings - Fork 0
Home
thisisnsh edited this page Aug 21, 2026
·
6 revisions
PlanX turns an AI coding agent's plan into a stored, versioned artifact you can read, annotate and approve outside the agent. It installs a skill into Codex and Claude Code, captures what they plan, and gives you a terminal review over the result. Approval refers to an exact version, and the version is what gets built.
PLAN → REVIEW → REVISE → EXECUTE → RESUME
| Page | What it covers |
|---|---|
| Installation | Installing the CLI, where the skills and the store land |
| The skill |
/planx and $planx, and the four branches they dispatch on |
| Page | What it covers |
|---|---|
| Planning | What the agent does on the plan branch, and the shape of a captured plan |
| Reviewing | The picker, the review screen, and every key on both |
| Feedback and edits | Line feedback, the plan-wide note, and direct edits |
| Versions and diffs | Moving between versions and reading the diff |
| Hand-offs |
s, the six exits, and what the receiving agent gets |
| Resuming a build |
ctrl+r back into the session that built a version |
| Custom agents |
planx defaults, and planning and building in different agents |
| Page | What it covers |
|---|---|
| CLI reference | Every command and flag, generated from the source |
| Troubleshooting | When something is not where you expect it |
| FAQ | Short answers, one question at a time |
Ask anything in Q&A — how a key works, whether a workflow is supported, why something behaves the way it does. Have something PlanX should do? Post it in Ideas, where it can be talked through before anyone builds it.
Start here
- Installation
- The skill
- Why review an agent's plan
- Plan mode vs. a reviewable plan
- Planning with Claude Code
- Planning with Codex
Using PlanX
Reference
Ask