By xonovex
Manage the full lifecycle of feature development: from acceptance criteria via Gherkin scenarios and structured plans, through isolated git worktrees, to code review with structured findings and PRs, and validation against criteria.
Draft: formalize acceptance criteria as Given-When-Then scenarios from user stories and designs, via three-amigos discovery and example mapping
Execution: validate a PR against its acceptance criteria, reporting pass/fail per scenario and flagging gaps
Commit and push changes in the current or a specified directory
Draft: approve a plan for execution — set status: approved after a final sanity check
Execution: resume work from an existing plan document with full context loading
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Monorepo for Xonovex AI agent tools, workflows, and skills
AI coding agents handle prompts, tools, and code changes. What they don't manage is the environment around them: sandbox isolation, model provider routing, terminal sessions, reproducible toolchains, and orchestration at scale.
Xonovex fills this gap. It currently supports Claude Code and OpenCode as agents, with sandboxing via bubblewrap and Docker, VM-level isolation via gVisor and Kata Containers, confidential computing via Confidential Containers (CoCo) with AMD SEV-SNP and Intel TDX, model routing through providers like Gemini, GLM, and GPT, workspace management with Git and Jujutsu, reproducible toolchains via Nix, and Kubernetes orchestration for running agents at scale.
The included skills are token-efficient, harness-neutral, and based on current research and best practices (Agent Skills spec, agentskills.io, agents.md).
npm install -g @xonovex/agent-cli-go
agent-cli run --agent claude --isolation bwrap --provider gemini
The sandbox is selected by three orthogonal axes — --isolation {none,bwrap,docker} ×
--provision {none,nix,command} × --network {host,none,proxy} — see
packages/agent/AGENTS.md.
# Install CRDs and deploy the operator
kubectl apply -k https://github.com/xonovex/platform//packages/agent/agent-operator-go/config/crd
kubectl apply -k https://github.com/xonovex/platform//packages/agent/agent-operator-go/config/default
# Create a provider and run an agent
kubectl create secret generic gemini-credentials --from-literal=api-key='your-key'
kubectl apply -f - <<EOF
apiVersion: agent.xonovex.com/v1alpha1
kind: AgentProvider
metadata:
name: gemini-provider
spec:
displayName: Google Gemini
authTokenSecretRef:
name: gemini-credentials
key: api-key
---
apiVersion: agent.xonovex.com/v1alpha1
kind: AgentRun
metadata:
name: review-code
spec:
harness:
type: claude
providerRef: gemini-provider
workspace:
type: git
repository:
url: https://github.com/org/repo.git
branch: main
prompt: "Review the codebase and suggest improvements"
EOF
Each skill is a separate plugin. Skills are applied automatically when the agent detects a relevant task — no explicit slash-command invocation needed. Skills also work cross-harness (Claude Code, OpenCode, etc.) since they follow the Agent Skills spec.
# Add the Xonovex plugin marketplace
claude plugin marketplace add xonovex/platform
npx claudepluginhub xonovex/platform --plugin xonovex-workflowXonovex ADR skills
Xonovex Docker skills
Xonovex Content skills
Xonovex Hono skills
Xonovex Android analytics skills
Harness-native ECC plugin for engineering teams - 67 agents, 279 skills, 94 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
The Claude Code knowledge system — 380+ skills, 182+ agents, 100+ commands, 40 hooks, 32 rules, and workflows.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Comprehensive C4 architecture documentation workflow with bottom-up code analysis, component synthesis, container mapping, and context diagram generation