[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78CWRSTRHM0QX3XQT85FBAW":3,"skill-same:01M78CWRSTRHM0QX3XQT85FBAW":77},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":34,"commitSha":35,"skillDirectory":36,"forksCount":14,"repositoryHtmlUrl":37,"repositoryHomepageUrl":38,"ownerHtmlUrl":39,"ownerBlogUrl":40,"ownerDescription":41,"updatedAt":42,"install":43,"resources":56},"01M78CWRSTRHM0QX3XQT85FBAW","jetson-optimize-memory",null,"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\u002FGPU frequency tuning.","Solizardking\u002Fskillhub-main","Solizardking","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F108096810?v=4","01M786KDB10AYV042N0F9BD5B5","Clawd","01M786NYYXSZK1D05WM8JQPC48",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: jetson-optimize-memory\ndescription: >-\n  Reclaim DRAM by disabling unused subsystems across MB1 BCT, MB2 BCT,\n  kernel reserved-memory, and SWIOTLB. Use for headless or no-camera\n  Jetson deployments; not for CPU\u002FGPU frequency tuning.\n\nversion: 0.0.1\nlicense: \"Apache-2.0\"\nargument-hint: \"headless | no-camera | swiotlb\"\nmetadata:\n  data-classification: public\n  author: \"Jetson Team\"\n  tags:\n    - memory\n    - dram\n    - carveout\n    - swiotlb\n  domain: memory\n---\n\n# jetson-optimize-memory\n\nMemory is reserved across four layers ordered by boot chronology\n(higher row = earlier in boot, closer to hardware):\n\n| Layer | Content | Key files |\n|---|---|---|\n| MB1 BCT | firmware carveouts | per-module misc DTS |\n| MB2 BCT | firmware loading + AST controls | per-module misc DTS |\n| Kernel DTS | reserved-memory and driver binding | per-module DTS |\n| SWIOTLB | DMA bounce pool size | `\u003Cmodule>.conf.common` (`CMDLINE_ADD`) |\n\n**Critical rules:**\n\n- Only the scenarios in **Scenario recipes** are validated. Refuse any\n  request to disable a carveout\u002Fcluster\u002Fnode not in that table.\n- **Zeroing a carveout requires both: disabling the cluster's loading\n  controls AND removing the AST that references it.**\n- **Emit explicit overrides for every cluster in the recipe**, regardless\n  of how the source `#ifdef`\u002F`#else` looks. Verify the merged binary.\n- **GPU SW stack must match the chip**: T234 -> nvgpu, T264 and later\n  -> OpenRM. Follow the shared derived-platform rule in\n  `target-platform-contract.md`; stop on mismatches instead of guessing.\n- Do not set SWIOTLB to 0 — some peripherals can't use the IOMMU.\n\n---\n\n## Scenario recipes\n\n| Keyword | MB1 BCT carveouts | MB2 BCT | Kernel DTS |\n|---|---|---|---|\n| `headless` | DCE-family (see chip table) | DCE `auxp_controls` + DCE AST(s) | `display@\u003Caddr>` (and `dce@\u003Caddr>` if exposed) → disabled |\n| `no-camera` | RCE\u002FVI\u002FISP-family | RCE `auxp_controls` (each instance) + RCE AST(s) | recommended: VI\u002FISP\u002FNVCSI → disabled |\n\n### Chip-specific carveouts\n\n| Scenario | T234 (Orin) | T264 (Thor) |\n|---|---|---|\n| `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` |\n| `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` |\n\n> **Post-boot:** `headless` → `sudo systemctl set-default multi-user.target`.\n\n---\n\n## MB1 BCT carveout overrides\n\nFile: `Linux_for_Tegra\u002Fbootloader\u002Fgeneric\u002FBCT\u002Ftegra\u003Cchip>-mb1-bct-misc-\u003Cmodule>.dts`\n(e.g. `tegra234-mb1-bct-misc-p3767-0000.dts` for Orin Nano,\n`tegra264-mb1-bct-misc-p3834-0008-p4071-0000.dts` for Thor).\n\nFor each carveout, add inside the existing `carveout` node:\n\n```dts\naux_info@\u003CCARVEOUT_NAME> {\n    pref_base = \u003C0x0 0x0>;\n    size      = \u003C0x0 0x0>;\n    alignment = \u003C0x0 0x0>;\n};\n```\n\n---\n\n## MB2 BCT cluster + AST overrides\n\nFile: `Linux_for_Tegra\u002Fbootloader\u002Fgeneric\u002FBCT\u002Ftegra\u003Cchip>-mb2-bct-misc-\u003Cmodule>.dts`\n(includes `tegra\u003Cchip>-mb2-bct-common.dtsi`).\n\nFor each target cluster:\n\n1. Override `auxp_controls@\u003Cindex>`:\n   ```dts\n   auxp_controls@\u003Cindex> {\n       enable_init    = \u003C0>;\n       enable_fw_load = \u003C0>;\n       enable_unhalt  = \u003C0>;\n   };\n   ```\n2. `\u002Fdelete-node\u002F auxp_ast_config@\u003Cidx>;`\n\nLook up indices in `common.dtsi`: `auxp_controls@N` carries a comment\nnaming its cluster; `auxp_ast_config@N` has `ast_region` children whose\n`carveout = \u003CCARVEOUT_…>;` lines identify the owner.\n\n---\n\n## Kernel DT reserved-memory\n\n```sh\nDTB=Linux_for_Tegra\u002Fkernel\u002Fdtb\u002F\u003Cplatform-dtb-name>.dtb\ndtc -I dtb -O dts -o \u002Ftmp\u002Fplatform.dts $DTB\n# edit: status = \"disabled\" on target nodes\ndtc -I dts -O dtb -o $DTB \u002Ftmp\u002Fplatform.dts\n```\n\n**Display** — disable `display@\u003Caddr>`, plus `dce@\u003Caddr>` if exposed as\na separate kernel node.\n\n**Camera** — under `host1x@\u003Caddr>`, disable whichever of `vi*` \u002F `isp*`\n\u002F `nvcsi` exist on the BSP (only emit present nodes):\n\nLocate the display controller node in the decompiled DTS and disable it.\nThe node's unit address is chip-specific — find it by `compatible` string\n(e.g. `nvidia,tegra234-display`) rather than hard-coding the address.\n\n```dts\nhost1x@\u003Caddr> {\n    vi0@\u003Caddr>   { status = \"disabled\"; };\n    vi1@\u003Caddr>   { status = \"disabled\"; };\n    isp@\u003Caddr>   { status = \"disabled\"; };\n    isp1@\u003Caddr>  { status = \"disabled\"; };\n    nvcsi@\u003Caddr> { status = \"disabled\"; };\n};\n```\n\n---\n\n## SWIOTLB DMA bounce pool\n\nThe NVIDIA IOMMU covers peripheral DMA, so SWIOTLB is rarely used.\nEdit `CMDLINE_ADD` (**never** `CMDLINE`) in `Linux_for_Tegra\u002F\u003Cmodule>.conf.common`:\n\n```sh\n# Total bytes = swiotlb_value × 2048; 4 MiB pool:\nCMDLINE_ADD=\"... swiotlb=2048\"\n```\n\n---\n\n## Override verification (mandatory)\n\nAfter every patched MB1\u002FMB2 BCT `.dts`, reproduce the BSP's compile +\ndecompile using the same `-D…` flags from `bct_flags.append(...)` in\n`bootloader\u002Ftegraflash_impl_t\u003Cchip>.py`:\n\n```sh\ngcc -E -nostdinc -x assembler-with-cpp \\\n    -DENABLE_\u003CFLAG_1> -DENABLE_\u003CFLAG_2> \\\n    -I bootloader -I bootloader\u002Fgeneric\u002FBCT \\\n    -o \u002Ftmp\u002Fcpp.dts \u003Cpatched-bct.dts>\ndtc -q -I dts -O dtb -o \u002Ftmp\u002Fcpp.dtb \u002Ftmp\u002Fcpp.dts\ndtc -q -I dtb -O dts \u002Ftmp\u002Fcpp.dtb | less\n```\n\nConfirm in the merged output:\n- Each zeroed `aux_info@\u003CNAME>` (or `aux_info@\u003Cid>U` post macro expansion)\n  has `size = \u003C0x0 0x0>` and `pref_base = \u003C0x0 0x0>`.\n- Each disabled `auxp_controls@\u003Cidx>` has all three `enable_*` fields `\u003C0>`.\n- Each `\u002Fdelete-node\u002F`'d `auxp_ast_config@\u003Cidx>` is absent.\n\n---\n\n## Verification (on booted target)\n\n```sh\nsudo cat \u002Fproc\u002Fiomem | grep -iE 'nv-reserved|cma|fb|carveout'\nls \u002Fproc\u002Fdevice-tree\u002Freserved-memory\u002F\ndmesg | grep -iE 'firmware|carveout|bpmp|reserved|fail|error' | head -20\nfree -m\n```\n\n| Scenario | Sysfs | dmesg grep |\n|---|---|---|\n| Display off | `ls \u002Fsys\u002Fclass\u002Fdrm\u002F` (empty) | `tegra-drm\\|nvdisplay\\|dce\\|host1x\\|fb0` |\n| Camera off | `ls \u002Fdev\u002Fvideo* 2>\u002Fdev\u002Fnull` (none) | `rce\\|nvcsi\\|tegra-camera\\|vi0\\|vi1\\|isp` |\n| SWIOTLB shrink | `cat \u002Fsys\u002Fkernel\u002Fdebug\u002Fswiotlb\u002Fio_tlb_nslabs` matches cmdline | `swiotlb` |\n\nFor SWIOTLB: `\u002Fproc\u002Fcmdline` must contain `swiotlb=\u003Cvalue>`, and\n`watch -n5 cat \u002Fsys\u002Fkernel\u002Fdebug\u002Fswiotlb\u002Fio_tlb_used` must stay under\n`io_tlb_nslabs` during full workload — if exceeded, restore original\n`CMDLINE_ADD` and re-flash `kernel-dtb`.\n\n## Purpose\n\nCut the unused DRAM carveouts that ship enabled in the reference BSP\nwhen a Jetson deployment skips display, camera, or other peripherals,\nfreeing the freed bytes for the application. Always edits the four\nlayers in boot order so an early-stage carveout never outranks a\nlater-stage shrink.\n\n## Prerequisites\n\n- Active target profile resolved per\n  `..\u002F..\u002Fcontext\u002Ftarget-platform-contract.md`.\n- BSP image extracted and source tree initialized\n  (`\u002Fjetson-init-image`, `\u002Fjetson-init-source` complete).\n- For headless \u002F no-camera recipes: confirm the workload truly does not\n  need display or camera.\n\n## Limitations\n\n- Only the validated recipes (`headless`, `no-camera`, `swiotlb`) are\n  exposed; ad-hoc subsystem disables outside the recipe set are\n  refused.\n- SWIOTLB shrink is bounded by peak in-flight DMA — exceeding the new\n  `io_tlb_nslabs` requires reverting the change.\n- BPMP-DTB edits land in the overlay tracker only after Customize +\n  Build + Deploy run; this skill does not flash on its own.\n\n## Troubleshooting\n\n- **Boot fails after MB1 BCT carveout disable** — restore the pristine\n  misc DTS and re-flash; the missing carveout is mandatory for the\n  active SoC.\n- **`io_tlb_used` exceeds `io_tlb_nslabs`** — revert `swiotlb=` in\n  `CMDLINE_ADD` and re-flash the kernel DTB partition.\n- **Reclaimed delta smaller than expected** — verify the recipe truly\n  matched the deployment (e.g. display still attached); use the\n  `dmesg | grep -iE 'firmware|carveout'` check in this file to confirm.\n- **Validation `dmesg` shows the disabled subsystem still probing** —\n  the change probably did not promote through to `bsp_image`; re-run\n  `\u002Fjetson-promote-image`.\n\n---\n","# jetson-optimize-memory\n\nMemory is reserved across four layers ordered by boot chronology\n(higher row = earlier in boot, closer to hardware):\n\n| Layer | Content | Key files |\n|---|---|---|\n| MB1 BCT | firmware carveouts | per-module misc DTS |\n| MB2 BCT | firmware loading + AST controls | per-module misc DTS |\n| Kernel DTS | reserved-memory and driver binding | per-module DTS |\n| SWIOTLB | DMA bounce pool size | `\u003Cmodule>.conf.common` (`CMDLINE_ADD`) |\n\n**Critical rules:**\n\n- Only the scenarios in **Scenario recipes** are validated. Refuse any\n  request to disable a carveout\u002Fcluster\u002Fnode not in that table.\n- **Zeroing a carveout requires both: disabling the cluster's loading\n  controls AND removing the AST that references it.**\n- **Emit explicit overrides for every cluster in the recipe**, regardless\n  of how the source `#ifdef`\u002F`#else` looks. Verify the merged binary.\n- **GPU SW stack must match the chip**: T234 -> nvgpu, T264 and later\n  -> OpenRM. Follow the shared derived-platform rule in\n  `target-platform-contract.md`; stop on mismatches instead of guessing.\n- Do not set SWIOTLB to 0 — some peripherals can't use the IOMMU.\n\n---\n\n## Scenario recipes\n\n| Keyword | MB1 BCT carveouts | MB2 BCT | Kernel DTS |\n|---|---|---|---|\n| `headless` | DCE-family (see chip table) | DCE `auxp_controls` + DCE AST(s) | `display@\u003Caddr>` (and `dce@\u003Caddr>` if exposed) → disabled |\n| `no-camera` | RCE\u002FVI\u002FISP-family | RCE `auxp_controls` (each instance) + RCE AST(s) | recommended: VI\u002FISP\u002FNVCSI → disabled |\n\n### Chip-specific carveouts\n\n| Scenario | T234 (Orin) | T264 (Thor) |\n|---|---|---|\n| `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` |\n| `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` |\n\n> **Post-boot:** `headless` → `sudo systemctl set-default multi-user.target`.\n\n---\n\n## MB1 BCT carveout overrides\n\nFile: `Linux_for_Tegra\u002Fbootloader\u002Fgeneric\u002FBCT\u002Ftegra\u003Cchip>-mb1-bct-misc-\u003Cmodule>.dts`\n(e.g. `tegra234-mb1-bct-misc-p3767-0000.dts` for Orin Nano,\n`tegra264-mb1-bct-misc-p3834-0008-p4071-0000.dts` for Thor).\n\nFor each carveout, add inside the existing `carveout` node:\n\n```dts\naux_info@\u003CCARVEOUT_NAME> {\n    pref_base = \u003C0x0 0x0>;\n    size      = \u003C0x0 0x0>;\n    alignment = \u003C0x0 0x0>;\n};\n```\n\n---\n\n## MB2 BCT cluster + AST overrides\n\nFile: `Linux_for_Tegra\u002Fbootloader\u002Fgeneric\u002FBCT\u002Ftegra\u003Cchip>-mb2-bct-misc-\u003Cmodule>.dts`\n(includes `tegra\u003Cchip>-mb2-bct-common.dtsi`).\n\nFor each target cluster:\n\n1. Override `auxp_controls@\u003Cindex>`:\n   ```dts\n   auxp_controls@\u003Cindex> {\n       enable_init    = \u003C0>;\n       enable_fw_load = \u003C0>;\n       enable_unhalt  = \u003C0>;\n   };\n   ```\n2. `\u002Fdelete-node\u002F auxp_ast_config@\u003Cidx>;`\n\nLook up indices in `common.dtsi`: `auxp_controls@N` carries a comment\nnaming its cluster; `auxp_ast_config@N` has `ast_region` children whose\n`carveout = \u003CCARVEOUT_…>;` lines identify the owner.\n\n---\n\n## Kernel DT reserved-memory\n\n```sh\nDTB=Linux_for_Tegra\u002Fkernel\u002Fdtb\u002F\u003Cplatform-dtb-name>.dtb\ndtc -I dtb -O dts -o \u002Ftmp\u002Fplatform.dts $DTB\n# edit: status = \"disabled\" on target nodes\ndtc -I dts -O dtb -o $DTB \u002Ftmp\u002Fplatform.dts\n```\n\n**Display** — disable `display@\u003Caddr>`, plus `dce@\u003Caddr>` if exposed as\na separate kernel node.\n\n**Camera** — under `host1x@\u003Caddr>`, disable whichever of `vi*` \u002F `isp*`\n\u002F `nvcsi` exist on the BSP (only emit present nodes):\n\nLocate the display controller node in the decompiled DTS and disable it.\nThe node's unit address is chip-specific — find it by `compatible` string\n(e.g. `nvidia,tegra234-display`) rather than hard-coding the address.\n\n```dts\nhost1x@\u003Caddr> {\n    vi0@\u003Caddr>   { status = \"disabled\"; };\n    vi1@\u003Caddr>   { status = \"disabled\"; };\n    isp@\u003Caddr>   { status = \"disabled\"; };\n    isp1@\u003Caddr>  { status = \"disabled\"; };\n    nvcsi@\u003Caddr> { status = \"disabled\"; };\n};\n```\n\n---\n\n## SWIOTLB DMA bounce pool\n\nThe NVIDIA IOMMU covers peripheral DMA, so SWIOTLB is rarely used.\nEdit `CMDLINE_ADD` (**never** `CMDLINE`) in `Linux_for_Tegra\u002F\u003Cmodule>.conf.common`:\n\n```sh\n# Total bytes = swiotlb_value × 2048; 4 MiB pool:\nCMDLINE_ADD=\"... swiotlb=2048\"\n```\n\n---\n\n## Override verification (mandatory)\n\nAfter every patched MB1\u002FMB2 BCT `.dts`, reproduce the BSP's compile +\ndecompile using the same `-D…` flags from `bct_flags.append(...)` in\n`bootloader\u002Ftegraflash_impl_t\u003Cchip>.py`:\n\n```sh\ngcc -E -nostdinc -x assembler-with-cpp \\\n    -DENABLE_\u003CFLAG_1> -DENABLE_\u003CFLAG_2> \\\n    -I bootloader -I bootloader\u002Fgeneric\u002FBCT \\\n    -o \u002Ftmp\u002Fcpp.dts \u003Cpatched-bct.dts>\ndtc -q -I dts -O dtb -o \u002Ftmp\u002Fcpp.dtb \u002Ftmp\u002Fcpp.dts\ndtc -q -I dtb -O dts \u002Ftmp\u002Fcpp.dtb | less\n```\n\nConfirm in the merged output:\n- Each zeroed `aux_info@\u003CNAME>` (or `aux_info@\u003Cid>U` post macro expansion)\n  has `size = \u003C0x0 0x0>` and `pref_base = \u003C0x0 0x0>`.\n- Each disabled `auxp_controls@\u003Cidx>` has all three `enable_*` fields `\u003C0>`.\n- Each `\u002Fdelete-node\u002F`'d `auxp_ast_config@\u003Cidx>` is absent.\n\n---\n\n## Verification (on booted target)\n\n```sh\nsudo cat \u002Fproc\u002Fiomem | grep -iE 'nv-reserved|cma|fb|carveout'\nls \u002Fproc\u002Fdevice-tree\u002Freserved-memory\u002F\ndmesg | grep -iE 'firmware|carveout|bpmp|reserved|fail|error' | head -20\nfree -m\n```\n\n| Scenario | Sysfs | dmesg grep |\n|---|---|---|\n| Display off | `ls \u002Fsys\u002Fclass\u002Fdrm\u002F` (empty) | `tegra-drm\\|nvdisplay\\|dce\\|host1x\\|fb0` |\n| Camera off | `ls \u002Fdev\u002Fvideo* 2>\u002Fdev\u002Fnull` (none) | `rce\\|nvcsi\\|tegra-camera\\|vi0\\|vi1\\|isp` |\n| SWIOTLB shrink | `cat \u002Fsys\u002Fkernel\u002Fdebug\u002Fswiotlb\u002Fio_tlb_nslabs` matches cmdline | `swiotlb` |\n\nFor SWIOTLB: `\u002Fproc\u002Fcmdline` must contain `swiotlb=\u003Cvalue>`, and\n`watch -n5 cat \u002Fsys\u002Fkernel\u002Fdebug\u002Fswiotlb\u002Fio_tlb_used` must stay under\n`io_tlb_nslabs` during full workload — if exceeded, restore original\n`CMDLINE_ADD` and re-flash `kernel-dtb`.\n\n## Purpose\n\nCut the unused DRAM carveouts that ship enabled in the reference BSP\nwhen a Jetson deployment skips display, camera, or other peripherals,\nfreeing the freed bytes for the application. Always edits the four\nlayers in boot order so an early-stage carveout never outranks a\nlater-stage shrink.\n\n## Prerequisites\n\n- Active target profile resolved per\n  `..\u002F..\u002Fcontext\u002Ftarget-platform-contract.md`.\n- BSP image extracted and source tree initialized\n  (`\u002Fjetson-init-image`, `\u002Fjetson-init-source` complete).\n- For headless \u002F no-camera recipes: confirm the workload truly does not\n  need display or camera.\n\n## Limitations\n\n- Only the validated recipes (`headless`, `no-camera`, `swiotlb`) are\n  exposed; ad-hoc subsystem disables outside the recipe set are\n  refused.\n- SWIOTLB shrink is bounded by peak in-flight DMA — exceeding the new\n  `io_tlb_nslabs` requires reverting the change.\n- BPMP-DTB edits land in the overlay tracker only after Customize +\n  Build + Deploy run; this skill does not flash on its own.\n\n## Troubleshooting\n\n- **Boot fails after MB1 BCT carveout disable** — restore the pristine\n  misc DTS and re-flash; the missing carveout is mandatory for the\n  active SoC.\n- **`io_tlb_used` exceeds `io_tlb_nslabs`** — revert `swiotlb=` in\n  `CMDLINE_ADD` and re-flash the kernel DTB partition.\n- **Reclaimed delta smaller than expected** — verify the recipe truly\n  matched the deployment (e.g. display still attached); use the\n  `dmesg | grep -iE 'firmware|carveout'` check in this file to confirm.\n- **Validation `dmesg` shows the disabled subsystem still probing** —\n  the change probably did not promote through to `bsp_image`; re-run\n  `\u002Fjetson-promote-image`.\n\n---\n",{"name":5,"license":15,"version":24,"metadata":25,"description":7,"argument-hint":33},"0.0.1",{"tags":26,"author":31,"domain":27,"data-classification":32},[27,28,29,30],"memory","dram","carveout","swiotlb","Jetson Team","public","headless | no-camera | swiotlb","main","be3de0244dc1a57e323d8160787ebb40fb64930e","skills\u002Fnvidia\u002Fjetson-optimize-memory","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main","https:\u002F\u002Fskills.x402.wtf","https:\u002F\u002Fgithub.com\u002FSolizardking","https:\u002F\u002Fmusebook.trade","Got a mile high and never came down","2026-10-08 23:20:13",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add Solizardking\u002Fskillhub-main --skill jetson-optimize-memory",{"tool":49,"command":50},"bunx","bunx skills add Solizardking\u002Fskillhub-main --skill jetson-optimize-memory",{"tool":52,"command":53},"pnpm","pnpm dlx skills add Solizardking\u002Fskillhub-main --skill jetson-optimize-memory","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002FSKILL.md",true,[57,63,67,71],{"path":58,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":61,"sourceHtmlUrl":62},"BENCHMARK.md","REFERENCE","参考文档",3839,"https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002FBENCHMARK.md",{"path":64,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":65,"sourceHtmlUrl":66},"evals\u002Fevals.json",4924,"https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002Fevals\u002Fevals.json",{"path":68,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":69,"sourceHtmlUrl":70},"skill-card.md",3704,"https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002Fskill-card.md",{"path":72,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":75,"sourceHtmlUrl":76},"skill.oms.sig","OTHER","其他",4597,"https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fnvidia\u002Fjetson-optimize-memory\u002Fskill.oms.sig",[78,86,93,100,107,115],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":19},"01M78E0TGCAMNPS207BA3G07S7","zkrouter","Self-hosted OpenAI-compatible LLM router — smart tier-based routing, ZK-stamped routing decisions, mode overrides, free OpenRouter at install via the Birth bot. Save 60-80% on inference costs by routing to the cheapest capable model across 12+ OpenRouter models.","MIT","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fzkrouter\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":19},"01M78E0R1Z8BDT6EDTHY17MTNB","zk","For custom ZK Solana programs and privacy-preserving applications to prevent double spending. Guide to integrate rent-free nullifier PDAs for double-spend prevention.","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fzk\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M78E0P3ZTZ1YWF8MD5MM23QE","zk-omni-messaging","Zero-knowledge omnichain messaging between Robinhood Chain and Solana using CheshireZkOmniMessenger (msgType 4) with nullifier anti-replay and the zk-omni-relayer service.","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fzk-omni-messaging\u002FSKILL.md",[],[],{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":19},"01M78E0JXZ2PNB1896VXAGDBWD","zerion-wallet-data","Use when the user asks to check a wallet's portfolio, token balances, DeFi positions, PnL, transaction history, or NFTs. Provides interpreted on-chain data across 40+ EVM chains and Solana via x402 pay-per-request or API key — no raw RPC decoding needed.","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fzerion-wallet-data\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":111,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M78E0G14FH9GRRTQYMKD5E59","youtube-clipper","YouTube 视频智能剪辑工具。下载视频和字幕，AI 分析生成精细章节（几分钟级别）， 用户选择片段后自动剪辑、翻译字幕为中英双语、烧录字幕到视频，并生成总结文案。 使用场景：当用户需要剪辑 YouTube 视频、生成短视频片段、制作双语字幕版本时。 关键词：视频剪辑、YouTube、字幕翻译、双语字幕、视频下载、clip video","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FSKILL.md",[],[],{"skillNo":116,"name":117,"nameZh":6,"description":118,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":82,"licenseFileUrl":6,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":19},"01M78E0C2ZYA99CFFDNB7ZMK5S","yield-optimization","Multi-chain yield optimization via yield.xyz (StakeKit). Discover 2,988+ yield opportunities across 75+ blockchains, build deposit\u002Fwithdrawal transactions with shell scripts, and sign them with a MoonPay wallet. Use when the user wants to earn yield, find the best lending or staking rate, enter\u002Fexit a position, or check portfolio balances.","https:\u002F\u002Fgithub.com\u002FSolizardking\u002Fskillhub-main\u002Fblob\u002Fmain\u002Fskills\u002Fyield-optimization\u002FSKILL.md",[],[]]