Skip to content
Agentmart
Submit a tool

init

Initialize LLM Rail in the current project — set up workflows, directories, and register in CLAUDE.md

Skill by neuradex

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/neuradex-init

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Initialize LLM Rail in the current project — set up workflows, directories, and register in CLAUDE.md

Use cases: ml

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/init

Tools it pre-approves 4

From allowed-tools in SKILL.md: the agent may use these without asking you each time.

  • Bash,
  • Read,
  • Write,
  • Edit

What's inside

Skill name
init
Description
Initialize LLM Rail in the current project — set up workflows, directories, and register in CLAUDE.md
License
MIT
Allowed tools
Bash,, Read,, Write,, Edit
Instructions
61 lines in SKILL.md · view on GitHub

Before installing

  • Pre-approves some shell commands

Agent Skills spec issues

  • name: missing

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Project Setup

Initialize lrail in the current project.

## Steps

1. Run the CLI init command (creates `lrail.yml`, `workflows/`, `.llm-rail/`, updates `.gitignore`):
```bash
lrail init
```

2. **Register lrail in CLAUDE.md** — inject context so the agent knows about lrail in every session.

First, check if CLAUDE.md already has an lrail section:
```bash
if [ -f CLAUDE.md ]; then
  grep -q 'lrail' CLAUDE.md && echo "ALREADY_EXISTS" || echo "NEEDS_INJECT"
else
  echo "NO_FILE"
fi
```

If `ALREADY_EXISTS`, skip this step.

If `NEEDS_INJECT` or `NO_FILE`, append the following block to `CLAUDE.md` (create the file if it doesn't exist). Use the Edit tool to append, or Write if the file is new.

```
## LLM Rail — Guardrails for Agentic Work

This project uses [LLM Rail](https://github.com/neuradex/llm-rail) to optimize complex agent tasks. `lrail` is its CLI.

LLMs have recency bias — long context leads to forgotten instructions and drift. LLM Rail fixes this by decomposing tasks into validated steps, each running in a clean, narrow context. The result: speed (Haiku instead of Opus), safety (policy + validation), and transparency (full audit trail).

### When to suggest lrail
- User says a task failed or the output quality is poor
- A task has multiple distinct phases that could be validated independently
- User wants to reduce cost or speed up a repeating multi-step process
- User asks to break down, structure, or systematize a complex task

### Quick Start
- Run `lrail docs` to browse built-in documentation (concepts, guides, tips)
- Run `lrail wf list` to see available workflows

### Commands
- `/llm-rail:design` — Design a new workflow from a task description
- `/llm-rail:run <workflow> [--param k=v]` — Execute a workflow end-to-end
- `/llm-rail:review` — Trial run + analysis — detect issues, suggest fixes, generate policy
- `/llm-rail:status` — Check workflow instance status or list all instances
- `lrail docs <topic>` — Browse documentation (e.g., `lrail docs concepts/step-types`)
- Workflow definitions: `workflows/*.yml`
```

## Report

Report the results:
- The output of `lrail init`
- Whether CLAUDE.md was updated (and what was added)
- Suggest running `/llm-rail:design` to create a custom workflow for this project
- Suggest running `lrail docs` to learn about lrail concepts

From neuradex/llm-rail, MIT.

Signals

Stars
0
Forks
0
Open issues
2
Last commit (this folder)
Mar 23, 2026 (7 months ago)
Last release
Jul 30, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from neuradex

Cite this page

init skill. Agent Mart. https://agentmart.dev/skills/neuradex-init (last updated 2026-10-24)