复制安装命令
用 Codex 或 Claude 安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它先审查 Skill 页面再帮你安装。
复制前请先查看来源、License 和安全提示。
SkillAtlasSkill 详情
jetson-optimize-memory
Official, NVIDIA-verified Agent Skills for Claude Code, Codex, and other coding agents.
审核状态：已审核Quality 80Security 62
项目 README
来源文件：README.md
NVIDIA Agent Skills
Official, NVIDIA-verified Agent Skills for Claude Code, Codex, and other coding agents.
📖 Docs: docs.nvidia.com/skills · 📺 Livestream: From Vulnerable to Verified · 📝 Blog: NVIDIA Verified Agent Skills: Capability Governance for AI Agents
Skills are portable instruction sets that teach AI agents how to use NVIDIA software optimally: Physical AI and robotics workflows, simulation, CUDA-X libraries, RAG and AI Blueprints, and platform tools. This repository is a catalog: skills are maintained in their respective product repos, and mirrored here daily via an automated sync pipeline. Skills are being added continuously, so check back for updates. We are building this infrastructure in the open, and contributions are welcome. See the Roadmap for what is planned next.
Quickstart
Install NVIDIA skills with the default skills CLI flow:
npx skills add nvidia/skills
The CLI runs through npx and prompts you to choose a skill and install destination. You do not need to clone this repo or copy skill folders by hand.
Requires a current
skillsCLI (v1.5.16 or newer). Installing vianpx skills@latest add nvidia/skillsalways uses the latest. On older CLIs (v1.5.15 and earlier), skills may install but not appear in Claude Code — see Troubleshooting.
The skill is available the next time your agent loads skills and encounters a relevant task. For example, ask your agent to "solve a linear programming problem with cuOpt" and the skill guides it through the cuOpt Python API. In Claude Code, run /reload-skills to load newly installed skills in your current session.
Install One Skill Without Prompts
Use this when you already know the skill name and want to skip prompts.
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --yes
Replace cuopt-numerical-optimization-api with any skill name from the Skill Catalog.
Install for a Specific Agent
Use --agent to target a specific AI coding agent. Initially, we'll support common client targets, expanding the list over time. For the full list of clients supported by the spec, see the skills CLI Supported Agents table.
Claude Code
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --agent claude-code
Codex
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --agent codex
Snowflake CoCo
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --agent cortex
Cursor
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --agent cursor
Kiro
npx skills add nvidia/skills --skill cuopt-numerical-optimization-api --agent kiro-cli
Use --agent more than once to install the same skill into multiple agents.
npx skills add nvidia/skills \
--skill cuopt-numerical-optimization-api \
--agent claude-code \
--agent codex \
--agent cursor \
--agent kiro-cli
Keep Skills Up to Date
New skills land continuously, and existing ones are revised, renamed, or consolidated as the catalog evolves. Refresh what you have installed with:
npx skills update
Run it interactively and the CLI also flags skills that were removed or merged upstream (for example, when several skills are consolidated into one) and offers to remove the stale local copies. Use npx skills list to see what is installed and npx skills check to preview what is out of date first.
Browse the Catalog
Use this when you want to see available NVIDIA skills before installing anything.
npx skills add nvidia/skills --list
For non-interactive installs, global installs, agent-specific installs, updates, removals, and fallback manual copying, see Advanced installation.
Skill Catalog
Getting Help & Contributing
Where to file an issue depends on what's broken:
- Skill content issues (a specific skill has a bug, missing functionality, or incorrect content) — file in the source repo for that product, using the per-product table below.
- Catalog issues (catalog README errors, sync workflow problems, distribution channels, signing/verification flow, docs in this repo) — file here using the catalog issue templates: Bug Report, Feature Request, or Documentation Request or Correction.
- Questions or general discussion — use Discussions. The issue tracker is reserved for bug reports, feature proposals with a design, and documentation issues.
- Security vulnerabilities — follow the disclosure process in SECURITY.md; do not open a public issue.
Per-product source repo links:
For issues with this catalog repo itself (README, structure, listing a new product): open an issue here.
Verifying Skills
Every published skill ships with a detached OMS signature (skill.oms.sig). The sync pipeline drops any skill missing the required artifacts before publishing, so every skill in the catalog carries:
SKILL.md— the skill instructions consumed by the agentskill-card.md— skill identity and governance cardskill.oms.sig— detached OMS signature (verifiable againstnv-agent-root-cert.pem)- A Tier-3 evaluation dataset — accepted at
evals/evals.json,evals/*.json,eval/*.json, orbenchmark/evals.json BENCHMARK.md— generated benchmark report capturing verifiable uplift data
Verify a skill against the NVIDIA trust anchor nv-agent-root-cert.pem:
pip install model-signing
model_signing verify certificate SKILL_DIR \
--signature SKILL_DIR/skill.oms.sig \
--certificate_chain nv-agent-root-cert.pem \
--ignore_unsigned_files
A successful verification confirms that the skill contents have not been modified since signing by NVIDIA.
See Verify Signed Agent Skills for signature layout, the trust pipeline, and policy options.
Roadmap
- ✅ Public skills catalog with NVIDIA-verified skills across multiple products
- ✅ Automated sync pipeline with skills mirrored from product repos daily
- ✅ Security scanning for all published skills covering instruction safety and supply-chain integrity
- ✅ Skills signing so every published skill carries a verifiable NVIDIA signature
- ✅ Skills universal evaluation criteria and task-specific criteria
- ✅ Skill Card with machine-readable metadata for identity, provenance, quality, and behavioral boundaries
- ✅ Sync-time compliance gates — signature drift detection and missing-artifact enforcement
- ✅ Syndication to external marketplaces — Skills.sh, Codex plugin, Claude Code plugin, ClawHub, Hermes Hub
- 🔲 Syndication to additional MCP hubs and partner channels
Repository Structure
NVIDIA/skills/
├── skills/ # NVIDIA-verified skills (count grows continuously),
│ │ synced from upstream product repos
│ ├── README.md # Browser-facing install guidance
│ ├── <product-prefix>-*/ # Flat layout — one dir per skill, product-prefixed
│ │ # e.g. aiq-*, cuopt-*, cupynumeric-*,
│ │ # dali-*, deepstream-*, dicom-*, digital-health-*,
│ │ # dynamo-*, earth2studio-*, holoscan-*, hsb-*,
│ │ # jetson-*, launch-nemo-rl, mcore-*,
│ │ # nemo-automodel-*, nemo-data-designer-plugin,
│ │ # nemo-evaluator-plugin, nemo-mbridge-* (20 skills),
│ │ # nemo-retriever, nemo-rl-* (4 skills),
│ │ # nemoclaw-user-guide, nemotron-*, nemotron-speech,
│ │ # nv-* (medical AI), physicsnemo-*, rag-*,
│ │ # skill-card-generator, tao-*, tilegym-*,
│ │ # vss-* (15 skills), accelerated-computing-cudf,
│ │ # cudaq-guide, portfolio-optimization
│ ├── omniverse-*/ # Physical AI — manually staged (see manual-components.yml)
│ └── physical-ai-*/ # Physical AI — manually staged
├── components.d/ # Product registry — one file per component, teams onboard here
│ ├── README.md # Schema and onboarding instructions
│ └── <product>.yml # one file per registered product
├── plugins/ # Packaged plugin distributions
│ └── nvidia-skills/ # Curated NVIDIA skills bundle (Claude Code, Codex)
├── plugins.d/ # Plugin build registry — config for `build-plugins.py`
│ ├── README.md
│ ├── _defaults.yml
│ └── nvidia-skills.yml
├── .claude-plugin/ # Claude Code marketplace metadata
│ └── marketplace.json
├── .agents/plugins/ # Agent marketplace metadata (other clients)
│ └── marketplace.json
├── docs/ # Long-form documentation (published via Fern)
│ ├── README.md # How to build the docs locally
│ ├── index.mdx
│ ├── advanced-install.mdx
│ ├── agent-skill-trust-pipeline.mdx
│ ├── release-checklist.mdx
│ ├── scanning-agent-skills.mdx
│ ├── signing-agent-skills.mdx
│ └── skill-cards.mdx
├── fern/ # Fern docs site configuration
├── .github/
│ ├── workflows/ # Sync pipeline, plugin validation, DCO check, author verify
│ └── scripts/ # regenerate-readme.sh, build-plugins.py,
│ # manual-components.yml (temp Physical AI catalog
│ # exception, removed after Computex 2026),
│ # marketplace/metadata.json (skill metadata sidecar)
├── nv-agent-root-cert.pem # Trust anchor for OMS signature verification
├── skills.sh.json # Skills.sh marketplace grouping config
├── CHANGELOG.md
├── CONTRIBUTING.md # Contribution guidelines
├── SECURITY.md # Security reporting policy
├── CODE_OF_CONDUCT.md # Community code of conduct
├── LICENSE-APACHE # Apache 2.0 (source code)
└── LICENSE-CC-BY-4.0 # CC BY 4.0 (documentation/skills)
Skills are maintained in their respective product repos (see the Source column in the Skill Catalog) and synced to this repo daily. Products only appear under skills/ after the sync pipeline confirms each skill carries:
skill.oms.sig— detached OMS-format signature (verifiable againstnv-agent-root-cert.pem)skill-card.md— skill identity and governance card- A Tier-3 evaluation dataset — accepted at
evals/evals.json,evals/*.json,eval/*.json, orbenchmark/evals.json
When evaluation runs produce a BENCHMARK.md, it ships alongside the skill so consumers can see verifiable benchmark uplift data.
Standards & Compatibility
This repository adheres to the Agent Skills specification:
- Skills are portable directories with a
SKILL.mdfile at their root. - Metadata uses YAML frontmatter with required
nameanddescriptionfields. - Skills follow a progressive disclosure model — lightweight metadata loads at startup, full instructions load on activation.
- Validate your skill using the
skills-refreference library.
License
Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
This code is dual-licensed with documentation/skills under the CC-BY-4.0 AND source code under Apache-2.0 license terms. The full license texts can be found in LICENSE-APACHE and LICENSE-CC-BY-4.0 respectively.
其他
高风险
- 来源需自行核对维护者身份。
- 包含脚本或命令调用，安装前请复核。
- 未检测到明显外部权限要求。
- 存在潜在风险命令，请谨慎安装。
- 扫描发现：3 条。
Codex — Git Clone 安装
- 安装前请先查看来源仓库和风险报告。
- 克隆仓库：git clone https://github.com/NVIDIA/skills.git
- 将 "skills/jetson-optimize-memory" 文件夹复制到 Codex 的 skills 目录中。
- 重启 Codex 让新的 skill 生效。
Codex — 手动复制安装
- 安装前请先查看来源仓库和风险报告。
- 从源仓库下载 SKILL.md 及相关文件。
- 在 Codex 的 skills 目录中创建新文件夹。
- 将所有 skill 文件复制到新文件夹中。
- 重启 Codex 让新的 skill 生效。
Claude Code — Git Clone 安装
- 安装前请先查看来源仓库和风险报告。
- 克隆仓库：git clone https://github.com/NVIDIA/skills.git
- 将 "skills/jetson-optimize-memory" 文件夹复制到 Claude Code 的 skills 目录中。
- 重启 Claude Code 让新的 skill 生效。
Claude Code — 手动复制安装
- 安装前请先查看来源仓库和风险报告。
- 从源仓库下载 SKILL.md 及相关文件。
- 在 Claude Code 的 skills 目录中创建新文件夹。
- 将所有 skill 文件复制到新文件夹中。
- 重启 Claude Code 让新的 skill 生效。
Cursor — Git Clone 安装
- 安装前请先查看来源仓库和风险报告。
- 克隆仓库：git clone https://github.com/NVIDIA/skills.git
- 将 "skills/jetson-optimize-memory" 文件夹复制到 Cursor 的 skills 目录中。
- 重启 Cursor 让新的 skill 生效。
Cursor — 手动复制安装
- 安装前请先查看来源仓库和风险报告。
- 从源仓库下载 SKILL.md 及相关文件。
- 在 Cursor 的 skills 目录中创建新文件夹。
- 将所有 skill 文件复制到新文件夹中。
- 重启 Cursor 让新的 skill 生效。
GitHub Copilot — Git Clone 安装
- 安装前请先查看来源仓库和风险报告。
- 克隆仓库：git clone https://github.com/NVIDIA/skills.git
- 将 "skills/jetson-optimize-memory" 文件夹复制到 GitHub Copilot 的 skills 目录中。
- 重启 GitHub Copilot 让新的 skill 生效。
GitHub Copilot — 手动复制安装
- 安装前请先查看来源仓库和风险报告。
- 从源仓库下载 SKILL.md 及相关文件。
- 在 GitHub Copilot 的 skills 目录中创建新文件夹。
- 将所有 skill 文件复制到新文件夹中。
- 重启 GitHub Copilot 让新的 skill 生效。
Windsurf — Git Clone 安装
- 安装前请先查看来源仓库和风险报告。
- 克隆仓库：git clone https://github.com/NVIDIA/skills.git
- 将 "skills/jetson-optimize-memory" 文件夹复制到 Windsurf 的 skills 目录中。
- 重启 Windsurf 让新的 skill 生效。
Windsurf — 手动复制安装
- 安装前请先查看来源仓库和风险报告。
- 从源仓库下载 SKILL.md 及相关文件。
- 在 Windsurf 的 skills 目录中创建新文件夹。
- 将所有 skill 文件复制到新文件夹中。
- 重启 Windsurf 让新的 skill 生效。
查看 SKILL.md 原文
name: jetson-optimize-memory
description: >-
Reclaim DRAM by disabling unused subsystems across MB1 BCT, MB2 BCT,
kernel reserved-memory, and SWIOTLB. Use for headless or no-camera
Jetson deployments; not for CPU/GPU frequency tuning.
version: 0.0.1
license: "Apache-2.0"
argument-hint: "headless | no-camera | swiotlb"
metadata:
data-classification: public
author: "Jetson Team"
tags:
- memory
- dram
- carveout
- swiotlb
domain: memoryjetson-optimize-memory
Memory is reserved across four layers ordered by boot chronology (higher row = earlier in boot, closer to hardware):
| Layer | Content | Key files |
|---|---|---|
| MB1 BCT | firmware carveouts | per-module misc DTS |
| MB2 BCT | firmware loading + AST controls | per-module misc DTS |
| Kernel DTS | reserved-memory and driver binding | per-module DTS |
| SWIOTLB | DMA bounce pool size | <module>.conf.common (CMDLINE_ADD) |
Critical rules:
- Only the scenarios in Scenario recipes are validated. Refuse any request to disable a carveout/cluster/node not in that table.
- Zeroing a carveout requires both: disabling the cluster's loading controls AND removing the AST that references it.
- Emit explicit overrides for every cluster in the recipe, regardless
of how the source
#ifdef/#elselooks. Verify the merged binary. - GPU SW stack must match the chip: T234 -> nvgpu, T264 and later
-> OpenRM. Follow the shared derived-platform rule in
target-platform-contract.md; stop on mismatches instead of guessing. - Do not set SWIOTLB to 0 — some peripherals can't use the IOMMU.
Scenario recipes
| Keyword | MB1 BCT carveouts | MB2 BCT | Kernel DTS |
|---|---|---|---|
headless | DCE-family (see chip table) | DCE auxp_controls + DCE AST(s) | display@<addr> (and dce@<addr> if exposed) → disabled |
no-camera | RCE/VI/ISP-family | RCE auxp_controls (each instance) + RCE AST(s) | recommended: VI/ISP/NVCSI → disabled |
Chip-specific carveouts
| Scenario | T234 (Orin) | T264 (Thor) |
|---|---|---|
headless | CARVEOUT_BPMP_DCE, CARVEOUT_DCE, CARVEOUT_DCE_TSEC, CARVEOUT_TSEC_DCE, CARVEOUT_DISP_EARLY_BOOT_FB | CARVEOUT_DCE, CARVEOUT_TSEC_DCE, CARVEOUT_HPSE_DCE, CARVEOUT_DISP_EARLY_BOOT_FB |
no-camera | CARVEOUT_RCE, CARVEOUT_CAMERA_TASKLIST | CARVEOUT_RCE, CARVEOUT_RCE1, CARVEOUT_RCE_RW, CARVEOUT_VI_TASKLIST, CARVEOUT_VI1_TASKLIST, CARVEOUT_ISP_TASKLIST, CARVEOUT_ISP1_TASKLIST |
Post-boot:
headless→sudo systemctl set-default multi-user.target.
MB1 BCT carveout overrides
File: Linux_for_Tegra/bootloader/generic/BCT/tegra<chip>-mb1-bct-misc-<module>.dts
(e.g. tegra234-mb1-bct-misc-p3767-0000.dts for Orin Nano,
tegra264-mb1-bct-misc-p3834-0008-p4071-0000.dts for Thor).
For each carveout, add inside the existing carveout node:
aux_info@<CARVEOUT_NAME> {
pref_base = <0x0 0x0>;
size = <0x0 0x0>;
alignment = <0x0 0x0>;
};
MB2 BCT cluster + AST overrides
File: Linux_for_Tegra/bootloader/generic/BCT/tegra<chip>-mb2-bct-misc-<module>.dts
(includes tegra<chip>-mb2-bct-common.dtsi).
For each target cluster:
- Override
auxp_controls@<index>:auxp_controls@<index> { enable_init = <0>; enable_fw_load = <0>; enable_unhalt = <0>; }; /delete-node/ auxp_ast_config@<idx>;
Look up indices in common.dtsi: auxp_controls@N carries a comment
naming its cluster; auxp_ast_config@N has ast_region children whose
carveout = <CARVEOUT_…>; lines identify the owner.
Kernel DT reserved-memory
DTB=Linux_for_Tegra/kernel/dtb/<platform-dtb-name>.dtb
dtc -I dtb -O dts -o /tmp/platform.dts $DTB
# edit: status = "disabled" on target nodes
dtc -I dts -O dtb -o $DTB /tmp/platform.dts
Display — disable display@<addr>, plus dce@<addr> if exposed as
a separate kernel node.
Camera — under host1x@<addr>, disable whichever of vi* / isp*
/ nvcsi exist on the BSP (only emit present nodes):
Locate the display controller node in the decompiled DTS and disable it.
The node's unit address is chip-specific — find it by compatible string
(e.g. nvidia,tegra234-display) rather than hard-coding the address.
host1x@<addr> {
vi0@<addr> { status = "disabled"; };
vi1@<addr> { status = "disabled"; };
isp@<addr> { status = "disabled"; };
isp1@<addr> { status = "disabled"; };
nvcsi@<addr> { status = "disabled"; };
};
SWIOTLB DMA bounce pool
The NVIDIA IOMMU covers peripheral DMA, so SWIOTLB is rarely used.
Edit CMDLINE_ADD (never CMDLINE) in Linux_for_Tegra/<module>.conf.common:
# Total bytes = swiotlb_value × 2048; 4 MiB pool:
CMDLINE_ADD="... swiotlb=2048"
Override verification (mandatory)
After every patched MB1/MB2 BCT .dts, reproduce the BSP's compile +
decompile using the same -D… flags from bct_flags.append(...) in
bootloader/tegraflash_impl_t<chip>.py:
gcc -E -nostdinc -x assembler-with-cpp \
-DENABLE_<FLAG_1> -DENABLE_<FLAG_2> \
-I bootloader -I bootloader/generic/BCT \
-o /tmp/cpp.dts <patched-bct.dts>
dtc -q -I dts -O dtb -o /tmp/cpp.dtb /tmp/cpp.dts
dtc -q -I dtb -O dts /tmp/cpp.dtb | less
Confirm in the merged output:
- Each zeroed
aux_info@<NAME>(oraux_info@<id>Upost macro expansion) hassize = <0x0 0x0>andpref_base = <0x0 0x0>. - Each disabled
auxp_controls@<idx>has all threeenable_*fields<0>. - Each
/delete-node/'dauxp_ast_config@<idx>is absent.
Verification (on booted target)
sudo cat /proc/iomem | grep -iE 'nv-reserved|cma|fb|carveout'
ls /proc/device-tree/reserved-memory/
dmesg | grep -iE 'firmware|carveout|bpmp|reserved|fail|error' | head -20
free -m
| Scenario | Sysfs | dmesg grep |
|---|---|---|
| Display off | ls /sys/class/drm/ (empty) | tegra-drm|nvdisplay|dce|host1x|fb0 |
| Camera off | ls /dev/video* 2>/dev/null (none) | rce|nvcsi|tegra-camera|vi0|vi1|isp |
| SWIOTLB shrink | cat /sys/kernel/debug/swiotlb/io_tlb_nslabs matches cmdline | swiotlb |
For SWIOTLB: /proc/cmdline must contain swiotlb=<value>, and
watch -n5 cat /sys/kernel/debug/swiotlb/io_tlb_used must stay under
io_tlb_nslabs during full workload — if exceeded, restore original
CMDLINE_ADD and re-flash kernel-dtb.
Purpose
Cut the unused DRAM carveouts that ship enabled in the reference BSP when a Jetson deployment skips display, camera, or other peripherals, freeing the freed bytes for the application. Always edits the four layers in boot order so an early-stage carveout never outranks a later-stage shrink.
Prerequisites
- Active target profile resolved per
../../context/target-platform-contract.md. - BSP image extracted and source tree initialized
(
/jetson-init-image,/jetson-init-sourcecomplete). - For headless / no-camera recipes: confirm the workload truly does not need display or camera.
Limitations
- Only the validated recipes (
headless,no-camera,swiotlb) are exposed; ad-hoc subsystem disables outside the recipe set are refused. - SWIOTLB shrink is bounded by peak in-flight DMA — exceeding the new
io_tlb_nslabsrequires reverting the change. - BPMP-DTB edits land in the overlay tracker only after Customize + Build + Deploy run; this skill does not flash on its own.
Troubleshooting
- Boot fails after MB1 BCT carveout disable — restore the pristine misc DTS and re-flash; the missing carveout is mandatory for the active SoC.
io_tlb_usedexceedsio_tlb_nslabs— revertswiotlb=inCMDLINE_ADDand re-flash the kernel DTB partition.- Reclaimed delta smaller than expected — verify the recipe truly
matched the deployment (e.g. display still attached); use the
dmesg | grep -iE 'firmware|carveout'check in this file to confirm. - Validation
dmesgshows the disabled subsystem still probing — the change probably did not promote through tobsp_image; re-run/jetson-promote-image.
评论 (0)
暂无评论，成为第一个评论者吧！