VLSIT_RTL_Generator_AI_Flow

AI-assisted RTL design and verification workflow: Claude Code slash commands and Codex CLI skills walk an agent from hardware spec through requirements, RTL, SVA assertions, testbench, verification, and documentation, with requirements traceability and human sign-off.

AgentNot statedActiveVerificationEmbedded

What it does and how it connects

VLSIT is an AI-assisted RTL design and verification workflow that organizes work from hardware specification through requirements, configuration, RTL, SVA assertions and testbench, verification and documentation. Claude Code slash commands and Codex CLI skills walk an agent through the phases, with a Requirement Traceability Matrix linking each requirement to its specification clause, reviewed SVA, test cases, DUT results and human sign-off.

RTL generation and verification generation are separate branches sharing an approved baseline: the verification branch builds SVA and testbenches from the approved verification contract without reading the RTL, and the flow gates on human review at requirements, contract, properties and final verification. The repository also ships a reusable RTL IP library, common_rtl, with FIFOs, AXI4, APB, AHB and TL-UL bridges, CRC variants and a HyperRAM controller, each with its own specification.

Tools used are Verilator for lint, Yosys for synthesis, and the installed simulator for SystemVerilog and SVA, on Linux with Claude Code or Codex CLI. The README states the analyzed snapshot has no LICENSE file and that use and distribution terms must be confirmed with the repository owner, so no license is recorded for this page. Saved design snapshots describe individual runs and do not replace reproducible regression, formal proof, STA or ASIC sign-off.

What engineers report

  • Followed the repo since late September, cloned it, and re-ran the downscaler snapshot with Icarus 12 and Verilator 5.018.
  • booley

    An agentic RTL IDE with sandboxed Developer Agents and typed EDA flows, rather than a slash-command workflow with human sign-off gates.

  • AMD Ross

    AMD's agentic assistant for Vivado and Vitis HLS, rather than a spec-to-RTL workflow with a reusable IP library.

  • eda-agent

    An MCP bridge that drives EDA tools from an agent, rather than a spec-to-RTL workflow with human sign-off gates.

  • agentic-hil

    Hardware-in-the-loop agent testing, rather than a spec-to-RTL workflow with SVA and a traceability matrix.

  • Rapidflare suite

    Rapidflare's internal eval practice for AI agents in electronics, rather than a spec-to-RTL workflow with human sign-off gates.

Sources

  1. 1READMEgithub.com
  2. 2GitHub repo statsapi.github.com
  3. 3GitHub issue 1github.com

Not verified

  • Pricing: none stated on the repo.
  • License: no license is recorded; the README states the analyzed snapshot has no LICENSE file and that use and distribution terms must be confirmed with the repository owner. The catalog license of source-available is unverified.
  • The issue 1 voice is a close English paraphrase of a post written in Vietnamese.
  • No GitHub releases exist, so latest release fields are null.
  • No Verilog error ids exist in the error map, so no related errors are recorded.

Last verified November 6, 2026. Listings are informational. The evidence score measures public, verifiable evidence and is not an endorsement or a compatibility claim.