me2resh/apexyard

Your AI engineering org, governed. Ship agent-built code safely to production.

What it solves

AI coding agents can rapidly prototype software, but they often lack the governance, safety, and tracking required for production-grade shipping. ApexYard provides a structured system to move agent-built code into production by enforcing a formal software development life cycle (SDLC), tracking technical decisions, and implementing strict merge gates.

How it works

ApexYard acts as a multi-project "ops repo" that you fork and use to manage a portfolio of other repositories. It uses a combination of plain-text configuration files, Markdown-based roles, and Bash shell hooks to govern AI agents (primarily Claude Code, but also others via adapters).

  • Governance Layer: Shell hooks mechanically block unsafe actions (e.g., preventing edits without a ticket or merging without approval).
  • Agent Personas: It defines specialized sub-agents (like Rex for code review and Hakim for security) and roles across six departments to provide structured context to the AI.
  • Skill Set: It provides a library of slash-commands (e.g., /setup, /decide, /code-review) to automate common SDLC tasks.
  • Registry: A central project registry allows a single ops repo to track the status and health of multiple managed projects.

Who it’s for

Developers and teams using AI coding agents (such as Claude Code, opencode, pi, or Codex) who need a professional governance framework to ensure agent-generated code is reviewed, documented, and safely deployed.

Highlights

  • Strict Merge Gates: Requires both an AI reviewer (Rex) and a named human to approve the exact commit before merging.
  • Ticket-First Workflow: Enforces that every change starts with a ticket to prevent ad-hoc, untracked edits.
  • Agent Decision Records: Preserves technical decisions in a structured format rather than losing them in chat history.
  • Tool Agnostic: While built for Claude Code, it uses Bash-based enforcement and adapters to support other AI coding tools.
  • Zero Runtime: Operates entirely through plain-text files and shell scripts with no hosted service or lock-in.

Related

  • Project
    notque/vexjoy-agent

    A toolkit that connects plain-English requests to specialized AI agents and workflows, using a four-layer system of agents, skills, hooks, and scripts to enforce evidence-based task completion.

  • Project
    griddynamics/rosetta

    An engineering governance framework for AI coding agents that provides shared, versioned instructions and structured workflows to ensure agents follow team standards and architecture.

  • Project
    karanb192/claude-code-hooks

    A marketplace of 20 plug‑ins (hooks) for Claude Code that enforce safety (blocking secret reads, dangerous shell commands, destructive git actions, config tampering) and add automation (auto‑stage, code formatting, session logging, stand‑up generation, bounty board, context‑cost leaderboard). Install via `/plugin marketplace add karanb192/claude-code-hooks` then `/plugin install <hook>@claude-code-hooks`. MIT‑licensed, fully tested, and mapped to the OWASP LLM Top 10 2026.

  • Project
    devchilll/scope

    A production-ready governance framework for enterprise AI agents that implements safety guardrails, compliance-as-code, IAM, and human-in-the-loop escalation.

  • Project
    garrytan/gstack

    A software factory framework that turns AI coding agents into a virtual engineering team with specialized roles for product planning, architectural review, QA, and security audits.