griddynamics/rosetta

An instruction layer for AI coding agent

What it solves

Rosetta addresses the lack of engineering governance and shared context for AI coding agents. While agents can write code, they often lack the specific architecture, standards, and discipline of a professional engineering team, leading to guesswork, inconsistent prompts across developers, and code that violates project constraints.

How it works

Rosetta acts as a governance layer that sits between your existing AI tools (like Claude Code, Cursor, or Copilot) and your codebase. It uses a layered instruction system—Core, Organization, and Project—to provide a versioned source of truth. These instructions are authored in Markdown and stored in Git, ensuring that standards flow from the organization level down to every project. At runtime, Rosetta orchestrates these instructions and adapts them to the specific agent being used, guiding the agent through structured workflows with built-in approval gates.

Who it’s for

Engineering teams and organizations that use AI coding agents and want to enforce consistent standards, capture senior engineer knowledge, and ensure that agents follow a professional software development lifecycle (SDLC) rather than just "generating and hoping."

Highlights

  • Layered Context: Propagates organization-wide standards and project-specific architecture to agents automatically.
  • Structured Workflows: Provides predefined flows for coding, requirements authoring, security reviews, and code analysis with explicit phases (Prepare $\rightarrow$ Research $\rightarrow$ Plan $\rightarrow$ Act $\rightarrow$ Validate).
  • Engineering Guardrails: Enforces safety checks for dangerous actions, sensitive data handling, and risk assessment.
  • Agent Agnostic: Works with MCP-compatible agents including Claude Code, Cursor, and Copilot.
  • Human-in-the-Loop: Implements mandatory approval gates and structured questioning to prevent agents from guessing and burning tokens on wrong implementations.

Related

  • Project
    addyosmani/agent-skills

    Agent Skills is a library of 25 Markdown‑based engineering workflows (spec, plan, build, test, review, ship, etc.) that can be installed into many AI coding agents via a simple CLI. It provides slash commands, auto‑build mode, specialist personas, and reference checklists to make AI‑generated code follow senior‑engineer best practices across the entire development lifecycle.

  • Project
    AmazingAng/old-coder

    A workflow and set of instructions for coding agents that replaces manual code review with a rigorous automated testing gauntlet and an evidence report.

  • Project
    tw93/Waza

    A library of structured engineering skills for AI agents that replaces generic AI outputs with disciplined habits like systematic debugging and design pressure-testing.

  • Project
    codeaholicguy/ai-devkit

    AI DevKit is a local npm‑based control plane that unifies configuration, monitoring, cross‑agent messaging, searchable SQLite memory, and composable engineering “skills” (requirements, design, testing, review, etc.) for many AI coding assistants such as Claude Code, Gemini CLI, Codex CLI, Cursor, and GitHub Copilot. It lets teams manage multiple agents from a single console, keep project conventions in a searchable store, and enforce a senior‑engineer workflow without modifying the underlying LLMs.

  • Project
    HoangNguyen0403/agent-skills-standard

    Agent Skills Standard (AGS) is an NPM‑distributed CLI that lets teams publish their coding‑standards, security rules, and SDLC workflow templates as tiny markdown “skills”. The CLI (`ags init` / `ags sync`) creates a `.skillsrc` config, downloads language‑specific skill packs (280 skills across 20+ stacks), and writes a hierarchical router (`AGENTS.md → _INDEX.md → SKILL.md`). AI coding assistants (Cursor, Claude Code, Copilot, Gemini, etc.) read this router and load only the relevant rules for the file they are editing, saving ~85 % of prompt tokens. An optional MCP server can enforce hooks, block secret‑leak edits, and provide runtime audit logs. The project is a genuine AI‑coding‑assistant utility, not a generic tutorial or list.