Samtalki Agentrepl.Jl
AgentREPL.jl is a Claude Code plugin for the Build phase that provides an STDIO MCP server for persistent, multi-session Julia REPL access to cut TTFX overhead for AI assistants.
by samtalki · github.com/samtalki/AgentREPL.jl
Attach a persistent Julia REPL to Claude via STDIO MCP so scientific and numeric code edits avoid repeated TTFX startup penalties.
Add it to Claude Code
Install the plugin in Claude Code. One command, paste-ready.
/plugin install samtalki-agentrepl.jl@samtalki/AgentREPL.jlBuilt to be called by your agent
Skillselion is itself an MCP server. Your agent can pull this entry and a paste-ready install config straight from the API - no copy-paste.
Retrieve this entry with skillselion.get_details("plugin:samtalki/AgentREPL.jl") and the paste-ready config with skillselion.get_install_config("plugin:samtalki/AgentREPL.jl").
What it does
AgentREPL.jl is a Claude Code–oriented plugin built around an MCP server that keeps Julia REPL processes alive for AI assistants instead of paying a fresh compilation and load tax on every prompt. Solo builders doing scientific computing, prototyping models, or maintaining Julia backends know TTFX and process churn can dominate short agent loops; this repo positions STDIO MCP, persistent sessions, multi-session isolation, and Revise.jl-friendly hot reloading as the fix. Once registered, Claude (and similar MCP-capable agents) execute Julia in a warm interpreter, which makes tight edit-run-debug cycles feasible for notebooks-turned-scripts, simulation code, and small CLI analytics tools. It is phase-specific build integrations: you install it when Julia is on your critical path, not for generic web SaaS frontend work. Complexity is intermediate because MCP configuration, Julia environments, and Revise workflows must be correct. The catalog treats it as a development-tooling bridge—closer to infrastructure for agent-driven Julia than to a packaged domain skill.
Highlights
- STDIO MCP server exposing persistent Julia REPL sessions to AI coding assistants
- Reduces time-to-first-execution (TTFX) versus cold-starting Julia each turn
- Multi-session support for parallel Julia workspaces
- Revise.jl-oriented hot-reloading workflow for iterative scientific code
- Single-plugin bundle (pluginCount: 1) aimed at data-science and scientific-computing stacks
Why builders use it
AI-assisted Julia work feels sluggish when every agent turn cold-starts the REPL and repeats compilation instead of reusing a warm session.
After you wire AgentREPL.jl into Claude Code, your assistant talks to long-lived Julia REPL sessions over MCP so iterative edits and Revise-backed reloads stay fast.
At a glance
- Type - Plugin in Development Tools.
- Adoption - 0 installs, 2 stars, 0 votes.
FAQ
Who is AgentREPL.jl for?
It is for developers and data scientists using Claude or other MCP assistants who write Julia daily and want persistent REPL sessions without TTFX pain.
When should I use AgentREPL.jl?
Use it during build when you are iteratively editing Julia code, simulations, or scientific pipelines and need the agent to execute in a warm interpreter.
How do I add AgentREPL.jl to my agent?
Install the samtalki/AgentREPL.jl Claude Code plugin, configure the STDIO MCP server in Claude Code, and point it at your Julia project with Revise.jl as needed.
Comments
Share how you use samtalki-agentrepl.jl, gotchas, or tips for other indie builders.
No comments yet - be the first to share how you use it.