Reins — AI skill for Claude Code
The control layer for autonomous AI agents.
How to install Reins
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open Richie2956/reins and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Reins does
The control layer for autonomous AI agents. Budget governor, policy engine, hash chained ledger and an evidence pack mapped to SOC 2, ISO 27001, NIST AI RMF and the EU AI Act. Claude Code plugin, Anthropic and OpenAI proxy, SDK adapters. MIT.
Alternatives in AI
- AI Research SKILLs — Visual Skills Pack for Obsidian: generate Canvas, Excalidraw, and Mermaid diagrams from text with Claude Code 5.7k ★
- Surface — Show ranked attack surface for a target from its recon manifest + hunt memory 3.8k ★
- ML Training Pack — Machine learning model training skills 1.6k ★
README
reins
**The control layer for autonomous AI agents.**
Reins is a budget governor and a governance ledger for agents that run on their own. The governor meters spend per agent, steps the model down as the budget falls, slows the heartbeat, puts idle loops to sleep and halts at zero. The ledger puts a policy check in front of every tool call, keeps a hash chained audit trail, and exports an evidence pack an auditor can read, mapped to SOC 2, ISO 27001, NIST AI RMF and the EU AI Act.
Open source under MIT. Works with Claude Code, the Anthropic and OpenAI SDKs, the Vercel AI SDK, and any runtime that can point at an OpenAI compatible base URL. Everything runs on your machine. Nothing is sent anywhere.
Sixty second start
npm i -g @reins/cli
cd your-agent-project
reins init # writes reins.yml and creates ~/.reins/reins.db
reins proxy # local metering proxy on http://localhost:4141
Point your runtime at the proxy. For the Anthropic SDK that is `ANTHROPIC_BASE_URL=http://localhost:4141`. For the OpenAI SDK, or anything that speaks the OpenAI chat shape, it is `OPENAI_BASE_URL=http://localhost:4141/v1`. Then run your agent as usual.
reins status # tier, spent, remaining, model, heartbeat, alive
reins evidence --from 2026-09-01 --to 2026-09-30 --out pack.html
That is the whole loop. The proxy meters every call, applies the tier, records usage to the ledger, and the evidence command turns the ledger into a self contained HTML file with the pack's hash in the footer.
Claude Code plugin
The plugin wires Reins into Claude Code hooks, so the policy engine sees every tool call before it runs and the ledger sees every file change.
/plugin marketplace add Richie2956/reins
/plugin install reins@reins
Or, from a checkout, `claude --plugin-dir ./packages/claude-code`. The plugin calls the `reins` binary, so install `@reins/cli` globally first. If the CLI is missing the hooks say so plainly and let the session
Related Skills
Claude Skills Governance Risk And Compliance
Claude Skills for Governance, Risk & Compliance (GRC): Expert-level compliance guidance for ISO 27001, SOC 2,
Global Blocks
Claims that survive crossing a boundary — addressable, versioned, hash-chained blocks for AI agents; portals n
SkillWorkflowEngine
Native macOS app for composing and executing AI-powered consulting workflows. Combine personas (WER) with skil
Threads MCP CLI
Threads MCP server and CLI for Claude Code and AI agents. 30 tools for posting, chained threads, carousels, re
Herdr Gamepad
Drive Herdr with a game controller. Patrol your AI agents, split panes and switch workspaces from the couch —
Art Of War
📜 Sun Tzu's Art of War applied to AI agent organization and task orchestration — 13 chapters mapped to practi
Related Agents
Compliance Mapper
Delegates to this agent when the user wants to map penetration-test findings to compliance frameworks — PCI DS
Sec Researcher
Read-only security research for ClaudeSec — threats, CVEs, attack vectors, and compliance frameworks (NIST, IS
Compliance Automation Specialist
Use this agent when you need to automate compliance processes for SOC 2, ISO 27001, GDPR, HIPAA, and other ent