Vibe Coding Discovermartin-loop
View on GitHubMake AI coding agents safe to scale autonomously: assign work, cap spend, enforce policy, verify output, roll back failures, learn from loops, and prove ROI across every repo.
MartinLoop is a governed runtime and CLI for AI coding agents. It adds budgets, iteration caps, edit scope, verifier checks, rollback evidence, and run receipts across Codex, Claude Code, and MCP hosts.
Use Cases
cap coding-agent spend with hard budgetsenforce allowed edit scope and stop conditionsrequire independent verifier evidence before completionroll back failed agent runsgenerate shareable run receipts and dossiersintegrate governed runs into MCP hostsgate GitHub Actions on agent budget and verificationcompare coding agents under one governed workflow
Built With
- Language
- TypeScript
- Frameworks
- TypeScript · Node.js · pnpm · MCP · Claude Code Plugin · GitHub Actions · Vitest · OpenAI-compatible APIs
Tags
ai-coding-agents · agent-governance · budget-enforcement · mcp · claude-code · codex · rollback · audit-trail · policy-as-code · observability · llmops · control-plane · verification · receipts · github-actions · cli