Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

foundry-rtl-workflow

Étoiles0

Forks0

Mis à jour14 juillet 2026 à 01:39

Use for foundry RTL tasks involving ChipCraftX RTL generation, model-route evidence, GitHub implementation issues, and handoff to EDA validation.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

rajshah4

rajshah4/semi-demo

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Ingénieurs électroniques (sauf informatique)Architecture et ingénierie·SOC 17-2072

Explorateur de fichiers

2 fichiers

SKILL.md

readonly

name	foundry-rtl-workflow
description	Use for foundry RTL tasks involving ChipCraftX RTL generation, model-route evidence, GitHub implementation issues, and handoff to EDA validation.

Foundry RTL Workflow

Core Rule

OpenHands orchestrates the workflow. Specialist models generate or summarize. EDA tools validate.

Model Roles

Sonnet-style orchestrator: planning, edits, final judgment
SambaNova Llama 3.3 70B: fast log triage and repeated small loops
ChipCraftX RTLGen 7B: Verilog/SystemVerilog generation and repair

Required Evidence

Every completed run should report:

event or issue that triggered the run
model route used
files changed
validation commands
pass/fail status
unresolved risks

Skill-Owned Scripts

scripts/chipcraftx_generate_rtl.py: calls the configured ChipCraftX Hugging Face provider route and writes generated RTL plus metadata under artifacts/chipcraftx/.

Run it from the repository root:

python3 skills/foundry-rtl-workflow/scripts/chipcraftx_generate_rtl.py \
  --summary "<issue title and concise RTL requirements>" \
  --max-new-tokens 700 \
  --retries 5

Use the generated RTL as a first draft only. The orchestrator still owns code review, integration into examples/sync_fifo/sync_fifo.sv, validation, PR evidence, and human-review framing.

Plus depuis ce dépôt

même dépôt

foundry-sidekick-launcher

rajshah4/semi-demo

Launch or describe the visible parent/child conversation workflow for foundry RTL requests. Use when an OpenHands automation receives an RTL request and needs a parent orchestrator plus read-only child scout conversations before implementation.

2026-07-140

foundry-model-routing

rajshah4/semi-demo

Use for Foundry IT requests that need model/tool routing, child automation handoff, air-gapped handling, RTL implementation, or EDA validation policy.

2026-07-130

foundry-eda-validation

rajshah4/semi-demo

Use for deterministic validation of foundry RTL changes with static checks and optional EDA tools such as Verilator, Icarus Verilog, and Yosys.

2026-07-130

foundry-context-sidekick

rajshah4/semi-demo

Read-only context scout for foundry RTL requests. Use before implementation when Codex/OpenHands needs to gather spec, repo, toolchain, and model-route context without editing code.

2026-07-130

name	foundry-rtl-workflow
description	Use for foundry RTL tasks involving ChipCraftX RTL generation, model-route evidence, GitHub implementation issues, and handoff to EDA validation.

Foundry RTL Workflow

Core Rule

OpenHands orchestrates the workflow. Specialist models generate or summarize. EDA tools validate.

Model Roles

Sonnet-style orchestrator: planning, edits, final judgment
SambaNova Llama 3.3 70B: fast log triage and repeated small loops
ChipCraftX RTLGen 7B: Verilog/SystemVerilog generation and repair

Required Evidence

Every completed run should report:

event or issue that triggered the run
model route used
files changed
validation commands
pass/fail status
unresolved risks

Skill-Owned Scripts

scripts/chipcraftx_generate_rtl.py: calls the configured ChipCraftX Hugging Face provider route and writes generated RTL plus metadata under artifacts/chipcraftx/.

Run it from the repository root:

python3 skills/foundry-rtl-workflow/scripts/chipcraftx_generate_rtl.py \
  --summary "<issue title and concise RTL requirements>" \
  --max-new-tokens 700 \
  --retries 5

Use the generated RTL as a first draft only. The orchestrator still owns code review, integration into examples/sync_fifo/sync_fifo.sv, validation, PR evidence, and human-review framing.