What problem does it solve?
This Skill removes the back-and-forth normally needed to turn a coding request into an executable implementation plan. It infers intent from the current conversation, decides whether to continue an existing plan or create a new one, and starts ralphex immediately without asking follow-up questions.
Core Features & Use Cases
- Automatic plan continuation or creation: Detects active work in docs/plans, reuses the most relevant unfinished plan when appropriate, or generates a new minimal plan from repository and conversation context.
- Execution mode selection: Chooses between native execution and SBX sandbox execution based on task risk, user intent, and environment availability.
- Operational safeguards: Applies fixed iteration limits, confirms launch through the progress log, and preserves host protection expectations when sandbox mode is selected.
- Use Case: When a user says ralphex go after an interrupted coding session, this Skill can recover the likely plan, restart autonomous execution, and return the exact progress file to watch.
Quick Start
Ask the assistant to use ralphex-auto to continue the current work or create a plan from the conversation and start execution immediately.