axrepl-input

Execute scripts in running Python, Node, Bash, and Zsh REPL sessions with completion detection and output capture.

Updated Mar 20, 2026
One-click install
npx skills add https://github.com/SpecterShell/axec --skill axrepl-input
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: axrepl-input
Source: https://github.com/SpecterShell/axec/tree/main/skills/axrepl-input
Command: npx skills add https://github.com/SpecterShell/axec --skill axrepl-input

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Send scripts to a running REPL session and reliably capture complete output without guessing timeouts or stopping conditions.

Core Features & Use Cases

  • Completion-aware input: wraps code to detect when the REPL has finished, for Python, Node, Bash, and Zsh REPLs.
  • Streaming output: streams results as soon as available and trims wrappers from the final output.
  • Session flexibility: targets a specific REPL session or uses the latest one, with optional driver override and structured JSON output.
  • Use cases: long-running computations, multi-line experimentation, and interactive debugging within persistent sessions.

Quick Start

Send a script to a running REPL session, ensure completion, and capture the full output with a single command.

Frequently Asked Questions about axrepl-input

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I capture complete output from a running Python or Node REPL without guessing timeouts?

Use completion-aware input to wrap code sent to a running Python or Node REPL, automatically detecting when execution finishes and capturing the full output without fixed timeouts. It trims wrapper artifacts and streams results as they arrive.

Can I send multi-line scripts to a persistent Bash or Zsh session for interactive debugging?

Yes, you can send multi-line scripts to a persistent Bash or Zsh REPL session for interactive debugging. The tool detects command completion automatically and streams the full output, making it ideal for long-running computations and multi-line experimentation.

Does this approach work with multiple active REPL sessions at once?

Yes, completion-aware input supports multiple active REPL sessions by allowing optional session targeting. You can direct scripts to a specific running session or default to the latest one, with optional driver overrides and structured JSON output for tracking.

What is the best way to automate code execution in a REPL and get structured JSON output?

Automate code execution in a REPL by applying driver-wrapped input that detects completion and captures output. This approach provides structured JSON output, streaming results as soon as they are available and trimming wrappers from the final response.

Why does my REPL automation stop running before a long-running computation finishes?

Standard REPL automation often stops early because it relies on fixed timeouts rather than completion detection. Completion-aware input wraps the code to detect exactly when the REPL finishes executing, reliably capturing the full output for long-running computations.

Do I need to install specific drivers to use completion-aware input for REPL sessions?

Yes, completion-aware input requires driver-wrapped input to function. You must configure the appropriate driver for your Python, Node, Bash, or Zsh REPL, though you can optionally override the driver when sending scripts to a target session.