ironclaw-reborn-orientation

Maps IronClaw repository architecture, crate ownership, and request flow for developer navigation.

12.6k|1.5k|Updated Feb 3, 2026
One-click install
npx skills add https://github.com/nearai/ironclaw --skill ironclaw-reborn-orientation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ironclaw-reborn-orientation
Source: https://github.com/nearai/ironclaw/tree/main/.claude/skills/ironclaw-reborn-orientation
Command: npx skills add https://github.com/nearai/ironclaw --skill ironclaw-reborn-orientation

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The IronClaw repository's own documentation drifts out of date, making it hard to know where a new feature, fix, prompt, or doc belongs, which crate owns a subsystem, or how a request flows through the system. This Skill provides a verified, up-to-date architectural map so you stop routing work into deleted crates or stale locations.

Core Features & Use Cases

  • Placement routing: Directs new features into the correct Reborn crate family (app, contracts, domains, events, extensions, kernel, lanes, loop, product, substrates) and new prompt templates into the owning crate via include_str!.
  • Request flow tracing: Documents the canonical path from the Vite frontend through ironclaw_webui, product contracts, TurnRunScheduler, RebornTurnRunExecutor, and the agent loop down to runtime lanes.
  • Drift detection: Explains that the v1 src/ monolith and crates like ironclaw_engine and ironclaw_gateway are deleted, and gives grep commands to verify any crate's real consumers.
  • Use Case: You are about to add a new endpoint and a doc tells you to edit ironclaw_gateway; this Skill tells you that crate no longer exists and routes you to the correct Reborn crate instead.

Quick Start

Ask the agent to orient you in the IronClaw repo and tell you which crate should own the feature you are about to build.

Frequently Asked Questions about ironclaw-reborn-orientation

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

FAQPage Schema
How do I find which crate owns a subsystem in the IronClaw repo?

Check the crate family layout under crates/ and the routing table in crates/AGENTS.md, then verify real consumers with grep -rl --include=Cargo.toml "<crate_name>" crates/ Cargo.toml. The family directories group all crates into app, contracts, domains, events, extensions, kernel, lanes, loop, product, and substrates.

Where should new features go in the IronClaw repository?

All new features go into the Reborn crates under crates/, with the binary built from crates/app/ironclaw_cli. The old v1 src/ monolith and crates like ironclaw_engine, ironclaw_tui, ironclaw_gateway, and ironclaw_oauth have been deleted.

How does a request flow through the IronClaw WebUI?

Requests flow from the Vite frontend in ironclaw_webui through route handlers to ProductSurface contracts, then to product views in ironclaw_assistant. Turn execution runs through SessionThreadService, TurnCoordinator, TurnRunScheduler, RebornTurnRunExecutor, PlannedDriver, and CanonicalAgentLoopExecutor down to runtime lanes.

What should I do when IronClaw docs contradict the actual codebase?

Treat the doc as drift and re-verify with the grep commands provided in the Skill rather than routing around crates that no longer exist. Any reference to ironclaw_engine, ironclaw_tui, ironclaw_gateway, or ironclaw_oauth indicates stale documentation.

What is the difference between .claude/skills and top-level skills in IronClaw?

The .claude/skills/ directory contains developer workflow skills like this orientation guide, while the top-level skills/ directory holds product runtime skills compiled into the shipping binaries. Editing skills/ changes the product itself, so consult .claude/rules/guidance-maintenance.md first.