From ddf787bd01320485e55d87c3fac8ae45bc65850a Mon Sep 17 00:00:00 2001 From: Sayali Kandarkar Date: Mon, 27 Apr 2026 14:23:05 -0700 Subject: [PATCH 1/3] Add Version column to README skills table Each component row gets a cell wrapped in HTML-comment markers ( ... ) that the sync workflow will patch with the upstream commit SHA and date. Initial values are placeholder em dashes; RAG Blueprint stays plain since it isn't in components.yml. Co-Authored-By: Claude Opus 4.7 (1M context) Signed-off-by: Sayali Kandarkar --- README.md | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/README.md b/README.md index ab49f8ba..7b2001fc 100644 --- a/README.md +++ b/README.md @@ -45,19 +45,19 @@ That's it — the skill activates automatically the next time your agent encount ## Available Skills -| Product | Description | Skills | Catalog | Source | -|---------|-------------|:------:|---------|--------| -| **CUDA-Q** | CUDA Quantum — onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications. | 1 | [`skills/CUDA-Q/`](skills/CUDA-Q) | [Source](https://github.com/NVIDIA/cuda-quantum/tree/main/.claude/skills) | -| **cuOpt** | GPU-accelerated optimization — vehicle routing, linear programming, quadratic programming, installation, server deployment, and developer tools. | 19 | [`skills/cuopt/`](skills/cuopt) | [Source](https://github.com/NVIDIA/cuopt/tree/main/skills) | -| **Megatron-Bridge** | Bridge between NeMo and Megatron — data processing, model conversion, and training utilities. | 9 | [`skills/Megatron-Bridge/`](skills/Megatron-Bridge) | [Source](https://github.com/NVIDIA-NeMo/Megatron-Bridge/tree/main/skills) | -| **Megatron-Core** | Large-scale distributed training — model parallelism, pipeline parallelism, and mixed precision. | 6 | [`skills/Megatron-Core/`](skills/Megatron-Core) | [Source](https://github.com/NVIDIA/Megatron-LM/tree/main/skills) | -| **Model-Optimizer** | Model optimization — quantization, sparsity, and distillation for efficient inference. | 5 | [`skills/Model-Optimizer/`](skills/Model-Optimizer) | [Source](https://github.com/NVIDIA/Model-Optimizer/tree/main/.claude/skills) | -| **NeMo Evaluator** | LLM evaluation — launch evaluations, access MLflow results, NeMo Evaluator Launcher assistant, and bring-your-own benchmarks. | 4 | [`skills/NeMo-Evaluator/`](skills/NeMo-Evaluator) | [Source](https://github.com/NVIDIA-NeMo/Evaluator/tree/main/packages/nemo-evaluator-launcher/.claude/skills) | -| **NeMo Gym** | RL training environments — add benchmarks, resources servers, agent wiring, and reward profiling. | 1 | [`skills/NeMo-Gym/`](skills/NeMo-Gym) | [Source](https://github.com/NVIDIA-NeMo/Gym/tree/main/.claude/skills) | -| **NemoClaw** | Secure agent sandboxing — run OpenClaw inside NVIDIA OpenShell with managed inference, policy management, remote deployment, sandbox monitoring, and contributor/maintainer workflows. | 21 | [`skills/NemoClaw/`](skills/NemoClaw) | [Source](https://github.com/NVIDIA/NemoClaw/tree/main/.agents/skills) | -| **Nemotron Voice Agent** | Real-time conversational AI — deploy speech-to-speech voice agents on Workstation, Jetson Thor, or Cloud NIMs. | 1 | [`skills/nemotron-voice-agent/`](skills/nemotron-voice-agent) | [Source](https://github.com/NVIDIA-AI-Blueprints/nemotron-voice-agent/tree/main/.agents/skills) | -| **RAG Blueprint** | RAG pipeline — deploy, configure, troubleshoot, and manage retrieval augmented generation with Docker Compose or Helm. | 1 | [`skills/rag/`](skills/rag) | [Source](https://github.com/NVIDIA-AI-Blueprints/rag/tree/main/skill-source/.agents/skills) | -| **TensorRT-LLM** | LLM inference optimization — model onboarding, performance analysis and optimization, kernel writing, CI diagnostics, code contribution, and codebase exploration. | 20 | [`skills/TensorRT-LLM/`](skills/TensorRT-LLM) | [Source](https://github.com/NVIDIA/TensorRT-LLM/tree/main/.claude/skills) | +| Product | Description | Skills | Catalog | Source | Version | +|---------|-------------|:------:|---------|--------|---------| +| **CUDA-Q** | CUDA Quantum — onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications. | 1 | [`skills/CUDA-Q/`](skills/CUDA-Q) | [Source](https://github.com/NVIDIA/cuda-quantum/tree/main/.claude/skills) | — | +| **cuOpt** | GPU-accelerated optimization — vehicle routing, linear programming, quadratic programming, installation, server deployment, and developer tools. | 19 | [`skills/cuopt/`](skills/cuopt) | [Source](https://github.com/NVIDIA/cuopt/tree/main/skills) | — | +| **Megatron-Bridge** | Bridge between NeMo and Megatron — data processing, model conversion, and training utilities. | 9 | [`skills/Megatron-Bridge/`](skills/Megatron-Bridge) | [Source](https://github.com/NVIDIA-NeMo/Megatron-Bridge/tree/main/skills) | — | +| **Megatron-Core** | Large-scale distributed training — model parallelism, pipeline parallelism, and mixed precision. | 6 | [`skills/Megatron-Core/`](skills/Megatron-Core) | [Source](https://github.com/NVIDIA/Megatron-LM/tree/main/skills) | — | +| **Model-Optimizer** | Model optimization — quantization, sparsity, and distillation for efficient inference. | 5 | [`skills/Model-Optimizer/`](skills/Model-Optimizer) | [Source](https://github.com/NVIDIA/Model-Optimizer/tree/main/.claude/skills) | — | +| **NeMo Evaluator** | LLM evaluation — launch evaluations, access MLflow results, NeMo Evaluator Launcher assistant, and bring-your-own benchmarks. | 4 | [`skills/NeMo-Evaluator/`](skills/NeMo-Evaluator) | [Source](https://github.com/NVIDIA-NeMo/Evaluator/tree/main/packages/nemo-evaluator-launcher/.claude/skills) | — | +| **NeMo Gym** | RL training environments — add benchmarks, resources servers, agent wiring, and reward profiling. | 1 | [`skills/NeMo-Gym/`](skills/NeMo-Gym) | [Source](https://github.com/NVIDIA-NeMo/Gym/tree/main/.claude/skills) | — | +| **NemoClaw** | Secure agent sandboxing — run OpenClaw inside NVIDIA OpenShell with managed inference, policy management, remote deployment, sandbox monitoring, and contributor/maintainer workflows. | 21 | [`skills/NemoClaw/`](skills/NemoClaw) | [Source](https://github.com/NVIDIA/NemoClaw/tree/main/.agents/skills) | — | +| **Nemotron Voice Agent** | Real-time conversational AI — deploy speech-to-speech voice agents on Workstation, Jetson Thor, or Cloud NIMs. | 1 | [`skills/nemotron-voice-agent/`](skills/nemotron-voice-agent) | [Source](https://github.com/NVIDIA-AI-Blueprints/nemotron-voice-agent/tree/main/.agents/skills) | — | +| **RAG Blueprint** | RAG pipeline — deploy, configure, troubleshoot, and manage retrieval augmented generation with Docker Compose or Helm. | 1 | [`skills/rag/`](skills/rag) | [Source](https://github.com/NVIDIA-AI-Blueprints/rag/tree/main/skill-source/.agents/skills) | — | +| **TensorRT-LLM** | LLM inference optimization — model onboarding, performance analysis and optimization, kernel writing, CI diagnostics, code contribution, and codebase exploration. | 20 | [`skills/TensorRT-LLM/`](skills/TensorRT-LLM) | [Source](https://github.com/NVIDIA/TensorRT-LLM/tree/main/.claude/skills) | — | --- From 6acbfeafa0330a5755585deb4fc575441b47d4f9 Mon Sep 17 00:00:00 2001 From: Sayali Kandarkar Date: Mon, 27 Apr 2026 15:45:12 -0700 Subject: [PATCH 2/3] Capture upstream SHA and commit date during sync After cloning each product repo, record the upstream HEAD short SHA, full SHA, committer date (YYYY-MM-DD), and repo name to /tmp/sync-versions.txt. The next step will consume this to patch the README Version column markers. Slug derivation matches the marker keys in README.md: lowercase the component name and replace spaces with dashes. Co-Authored-By: Claude Opus 4.7 (1M context) Signed-off-by: Sayali Kandarkar --- .github/workflows/sync-skills.yml | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/.github/workflows/sync-skills.yml b/.github/workflows/sync-skills.yml index 10d937b5..07a7afc5 100644 --- a/.github/workflows/sync-skills.yml +++ b/.github/workflows/sync-skills.yml @@ -45,6 +45,7 @@ jobs: truncate -s 0 /tmp/rsynced-components.txt truncate -s 0 /tmp/changed-components.txt truncate -s 0 /tmp/failed-components.txt + truncate -s 0 /tmp/sync-versions.txt # Configure git to use the token via credential helper # so it never appears in clone URLs or error output @@ -69,6 +70,13 @@ jobs: continue fi + # Capture upstream HEAD info for the README Version column + upstream_sha=$(git -C "$tmp_dir" rev-parse HEAD) + upstream_short_sha=$(git -C "$tmp_dir" rev-parse --short HEAD) + upstream_date=$(git -C "$tmp_dir" show -s --format=%cs HEAD) + slug=$(echo "$name" | tr 'A-Z ' 'a-z-') + echo "$slug|$upstream_short_sha|$upstream_sha|$upstream_date|$repo" >> /tmp/sync-versions.txt + # Collect all skill paths for sparse checkout skill_count=$(yq -r ".components[$i].skills | length" components.yml) sparse_paths="" From 2f0e909442dbdafc3eb08e04c05cc10b2c647dd1 Mon Sep 17 00:00:00 2001 From: Sayali Kandarkar Date: Mon, 27 Apr 2026 15:45:39 -0700 Subject: [PATCH 3/3] Patch README version markers after each sync MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit New workflow step reads /tmp/sync-versions.txt and sed-replaces each component's marker block in README.md with a linked short SHA and the upstream committer date, e.g. [\`abc1234\`](https://github.com/.../commit/...) · 2026-04-27. Runs after "Fail if nothing synced" so total failures don't dirty README. Components without markers (e.g. RAG Blueprint, not in components.yml) are silently skipped — sed pattern just doesn't match. Uses | as the sed delimiter so URL slashes don't need escaping. Co-Authored-By: Claude Opus 4.7 (1M context) Signed-off-by: Sayali Kandarkar --- .github/workflows/sync-skills.yml | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/.github/workflows/sync-skills.yml b/.github/workflows/sync-skills.yml index 07a7afc5..21f46205 100644 --- a/.github/workflows/sync-skills.yml +++ b/.github/workflows/sync-skills.yml @@ -130,6 +130,20 @@ jobs: exit 1 fi + - name: Update README version markers + run: | + set -euo pipefail + if [ ! -s /tmp/sync-versions.txt ]; then + echo "No version data captured — skipping README update" + exit 0 + fi + while IFS='|' read -r slug short_sha full_sha commit_date repo; do + cell="[\`${short_sha}\`](https://github.com/${repo}/commit/${full_sha}) · ${commit_date}" + # Sed delimiter is | because URLs contain / and cell content has no | + sed -i "s|.*|${cell}|" README.md + done < /tmp/sync-versions.txt + echo "README version markers updated for $(wc -l < /tmp/sync-versions.txt) components" + - name: Build sync summary id: summary run: |