[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-doca-sha":3,"mdc--gegr0v-key":34,"related-repo-nvidia-doca-sha":1328,"related-org-nvidia-doca-sha":1434},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"doca-sha","offload SHA hashing to DOCA accelerators","Use this skill when the user is doing hands-on DOCA SHA programming — offloading SHA-1, SHA-256, or SHA-512 hashing onto a BlueField DPU or ConnectX accelerator, picking between one-shot `doca_sha_task_hash` and incremental `doca_sha_task_partial_hash`, querying `doca_sha_cap_*` for algorithm support and min destination \u002F max source buffer sizes, setting source \u002F destination `doca_mmap` permissions, or decoding DOCA_ERROR_* returns from the SHA API. Trigger even when the user does not explicitly mention \"DOCA SHA\" or \"doca_sha_task\" — typical implicit phrasings include \"hash a multi-GiB file on the DPU\", \"offload SHA-256 to the BlueField\", \"streaming hash over chunks\", \"partial hash returns BAD_STATE\", \"destination buffer too small for digest\", or \"is SHA-512 available on this card\". Refuse and route elsewhere for general cryptographic-hash theory (collision resistance, SHA-3 selection), other DOCA crypto libraries (AES-GCM, Compress, DMA), or DOCA install \u002F BFB bring-up — those belong to other skills.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Cryptography","cryptography","tag",{"name":17,"slug":18,"type":15},"Hardware","hardware",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Engineering","engineering",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-20T06:24:04.311416","Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fdoca-sha","---\nlicense: Apache-2.0\nname: doca-sha\ndescription: >\n  Use this skill when the user is doing hands-on DOCA SHA programming —\n  offloading SHA-1, SHA-256, or SHA-512 hashing onto a BlueField DPU or\n  ConnectX accelerator, picking between one-shot `doca_sha_task_hash` and\n  incremental `doca_sha_task_partial_hash`, querying `doca_sha_cap_*`\n  for algorithm support and min destination \u002F max source buffer sizes, setting\n  source \u002F destination `doca_mmap` permissions, or decoding DOCA_ERROR_*\n  returns from the SHA API. Trigger even when the user does not explicitly\n  mention \"DOCA SHA\" or \"doca_sha_task\" — typical implicit phrasings\n  include \"hash a multi-GiB file on the DPU\", \"offload SHA-256 to the\n  BlueField\", \"streaming hash over chunks\", \"partial hash returns\n  BAD_STATE\", \"destination buffer too small for digest\", or \"is SHA-512\n  available on this card\". Refuse and route elsewhere for general\n  cryptographic-hash theory (collision resistance, SHA-3 selection), other\n  DOCA crypto libraries (AES-GCM, Compress, DMA), or DOCA install \u002F BFB\n  bring-up — those belong to other skills.\nmetadata:\n  kind: library\ncompatibility: >\n  Requires DOCA SDK installed at \u002Fopt\u002Fmellanox\u002Fdoca on Linux (Ubuntu\n  22.04\u002F24.04 or RHEL\u002FSLES) with a BlueField DPU or ConnectX NIC attached.\n  Reads the user's local install via `pkg-config doca-sha` and inspects\n  \u002Fopt\u002Fmellanox\u002Fdoca\u002F{lib,include,samples,applications}.\n---\n\n# DOCA SHA\n\n**Where to start:** This skill assumes DOCA is already installed and\nthe user is doing **hands-on SHA-acceleration work** on a BlueField\n\u002F ConnectX \u002F host with DOCA. Open [`TASKS.md`](TASKS.md) if the user\nwants to *do* something (configure \u002F build \u002F modify \u002F run \u002F test \u002F\ndebug); open [`CAPABILITIES.md`](CAPABILITIES.md) when the question\nis *what can DOCA SHA express* on this version. If the user has not\ninstalled DOCA yet, route to\n[`doca-setup`](..\u002F..\u002Fdoca-setup\u002FSKILL.md) first. If the user is\nasking *\"should I even use the accelerator for this hash?\"*, the\npath-selection rule in\n[`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)\nis the first stop.\n\n## Example questions this skill answers well\n\nThe CLASSES of DOCA SHA questions this skill is built to answer,\neach with one worked example. The agent should treat the *class* as\nthe load-bearing piece — the worked example is a single instance.\n\n- **\"Should I offload this hash to DOCA SHA, or just compute it on\n  the CPU?\"** — worked example: *\"I am verifying file integrity on\n  a 4 GiB image; is doca-sha worth the setup vs OpenSSL on the\n  CPU?\"*. Answered by the path-selection table in\n  [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)\n  + the *\"when NOT to use doca-sha\"* bullets in\n  [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy).\n- **\"Does my device support the SHA algorithm I want?\"** — worked\n  example: *\"is SHA-256 in the accelerator on this BlueField, and\n  what is the minimum destination buffer size for it?\"*. Answered\n  by the algorithm + buffer-sizing capability-query rule\n  (`doca_sha_cap_task_hash_get_supported(devinfo, algorithm)`\n  for the one-shot path; `_task_partial_hash_get_supported(devinfo, algorithm)`\n  for the partial-hash path; `doca_sha_cap_get_min_dst_buf_size`,\n  `doca_sha_cap_get_max_src_buf_size`) in\n  [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)\n  + the discovery step in\n  [`TASKS.md ## configure`](TASKS.md#configure).\n- **\"How do I pick between the one-shot and the partial \u002F incremental\n  hash task?\"** — worked example: *\"my input is 1 GiB and the device\n  cap says max source buffer is 64 MiB\"*. Answered by the\n  one-shot-vs-partial table in\n  [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)\n  + the task-config workflow in\n  [`TASKS.md ## modify`](TASKS.md#modify).\n- **\"What permissions does the source \u002F destination mmap need?\"** —\n  worked example: *\"my `doca_sha_task_hash` returns\n  `DOCA_ERROR_NOT_PERMITTED`\"*. Answered by the permission matrix\n  in [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy)\n  + the mmap-set-permissions checklist in\n  [`TASKS.md ## test`](TASKS.md#test).\n- **\"Is this DOCA SHA API available on my installed DOCA version?\"**\n  — worked example: *\"is `doca_sha_task_partial_hash` in the DOCA\n  I have installed?\"*. Answered by the version-compatibility\n  overlay in\n  [`CAPABILITIES.md ## Version compatibility`](CAPABILITIES.md#version-compatibility),\n  which cross-links the canonical detection chain in\n  [`doca-version`](..\u002F..\u002Fdoca-version\u002FSKILL.md) and adds the\n  SHA-specific *\"discover, do not assume\"* bullets.\n- **\"What does this `DOCA_ERROR_*` from a SHA call mean and which\n  layer caused it?\"** — worked example: *\"`DOCA_ERROR_INVALID_VALUE`\n  on `doca_sha_task_hash_alloc_init`\"*. Answered by the SHA overlay\n  on the cross-library taxonomy in\n  [`CAPABILITIES.md ## Error taxonomy`](CAPABILITIES.md#error-taxonomy)\n  + the layered ladder in\n  [`TASKS.md ## debug`](TASKS.md#debug) that escalates to\n  [`doca-debug`](..\u002F..\u002Fdoca-debug\u002FSKILL.md).\n\n## Audience\n\nThis skill serves **external developers building applications that\nconsume the DOCA SHA library** — i.e., users whose code calls\n`doca_sha_*` (directly in C\u002FC++, or through FFI\u002Fbindings from\nanother language) to offload SHA hashing onto a BlueField DPU or\nConnectX accelerator. It is *not* for NVIDIA developers contributing\nto DOCA SHA itself.\n\n**Language scope.** DOCA SHA ships as a C library with `pkg-config`\nmodule name `doca-sha`. The shipped samples are written in C. C and\nC++ consumers are the canonical case and the worked examples in\n`TASKS.md` assume that path. Other-language consumers (Rust, Go,\nPython, …) consume the same `*.so` through FFI or language-specific\nbindings; the skill's contribution in that case is to keep the\nlifecycle, capability-discovery, permission, error-taxonomy, and\none-shot-vs-partial guidance language-neutral, and to route the\nagent to the public C ABI as the authoritative surface that any\nwrapper will eventually call.\n\n## When to load this skill\n\nLoad this skill when the user is doing hands-on DOCA SHA work, in\nany language. Concretely:\n\n- Initializing a `doca_sha` context on a `doca_dev` and configuring\n  at least one task type (`doca_sha_task_hash` and\u002For\n  `doca_sha_task_partial_hash`) before `doca_ctx_start()`.\n- Choosing between the **one-shot** task (`doca_sha_task_hash` —\n  input fits in a single source buffer, output digest lands in a\n  single destination buffer) and the **partial \u002F incremental** task\n  (`doca_sha_task_partial_hash` — input streamed in chunks, finalized\n  separately) for the user's data shape.\n- Setting permissions on `doca_mmap` correctly for the source buffer\n  (`DOCA_ACCESS_FLAG_LOCAL_READ_ONLY` at minimum) and the destination\n  buffer (`DOCA_ACCESS_FLAG_LOCAL_READ_WRITE`).\n- Sizing the destination buffer against\n  `doca_sha_cap_get_min_dst_buf_size(devinfo, algorithm)` and\n  the source buffer against\n  `doca_sha_cap_get_max_src_buf_size(devinfo)`.\n- Checking which SHA algorithm enums\n  (`DOCA_SHA_ALGORITHM_SHA1`, `DOCA_SHA_ALGORITHM_SHA256`,\n  `DOCA_SHA_ALGORITHM_SHA512`) the active device's accelerator\n  advertises, via\n  `doca_sha_cap_task_hash_get_supported(devinfo, algorithm)` and\n  `doca_sha_cap_task_partial_hash_get_supported(devinfo, algorithm)` —\n  both fold task-support and algorithm-support into one call.\n- Validating a digest against a published test vector before pushing\n  bulk input through the accelerator.\n- Debugging a `DOCA_ERROR_*` returned from a SHA call (lifecycle vs.\n  buffer-sizing vs. permission vs. unsupported-algorithm) and the\n  task-completion event on the progress engine.\n- Designing or extending non-C bindings (Rust, Go, Python, …) that\n  wrap the SHA C ABI — for the lifecycle, permission, capability,\n  and one-shot-vs-partial rules the wrapper must honor.\n\nDo **not** load this skill for general DOCA orientation, install of\nDOCA itself, non-SHA hashing libraries on CPU (use OpenSSL or\nsimilar), or other DOCA libraries. For those, use\n[`doca-public-knowledge-map`](..\u002F..\u002Fdoca-public-knowledge-map\u002FSKILL.md).\n\n## What this skill provides\n\nThis is a **thin loader**. The body keeps only the orientation\nneeded to pick the right next file. The substantive SHA-specific\nmaterial lives in two companion files:\n\n- `CAPABILITIES.md` — what DOCA SHA can express on this version:\n  the two task types (one-shot hash and partial \u002F incremental\n  hash), the three algorithm enums, the capability-query surface\n  (`doca_sha_cap_*` for algorithm support and buffer sizing), the\n  SHA error taxonomy (mapped onto the cross-library `DOCA_ERROR_*`\n  set), the observability surface (per-task completion events on\n  the progress engine), the safety policy that gates source \u002F\n  destination mmap permission decisions, and the path-selection\n  rule (when to use doca-sha versus a CPU hash or a different\n  DOCA crypto library).\n- `TASKS.md` — step-by-step workflows for the six in-scope SHA\n  verbs: `configure`, `build`, `modify`, `run`, `test`, `debug`.\n  Plus a `Deferred task verbs` block that points out-of-scope\n  questions at the right next skill.\n\nThe skill assumes a host or BlueField where DOCA is already\ninstalled at the standard location and the user has the privileges\ntheir public install profile expects. It does not cover installing\nDOCA — that path goes through\n[`doca-setup`](..\u002F..\u002Fdoca-setup\u002FSKILL.md).\n\n## What this skill deliberately does not ship\n\nThis skill is **agent guidance**, not a samples or templates\nbundle. To keep the boundary clean, it deliberately does not\ncontain — and pull requests should not add:\n\n- **Pre-written DOCA SHA application source code, in any\n  language.** The verified SHA source code is the shipped C samples\n  at `\u002Fopt\u002Fmellanox\u002Fdoca\u002Fsamples\u002Fdoca_sha\u002F`, plus the File Integrity\n  reference application linked from the public DOCA SHA guide. The\n  agent's job is to route the user to those files and prescribe a\n  minimum-diff modification on them via the universal\n  modify-a-sample workflow in\n  [`doca-programming-guide`](..\u002F..\u002Fdoca-programming-guide\u002FSKILL.md),\n  layered with the SHA-specific overrides in\n  [`TASKS.md ## modify`](TASKS.md#modify).\n- **Pre-computed digest tables for arbitrary inputs.** The skill\n  tells the agent to use a *published* test vector (e.g. the NIST\n  SHA test vectors for the empty string, \"abc\", and the\n  million-`a` input) as the known-vector smoke; it does not ship a\n  vector bank of its own.\n- **Standalone build manifests** (`meson.build`, `CMakeLists.txt`,\n  `Cargo.toml`, …) parked inside the skill. The agent constructs\n  the build manifest *in the user's project directory* against the\n  user's installed DOCA, where `pkg-config --modversion doca-sha`\n  is the source of truth.\n- **A `samples\u002F`, `bindings\u002F`, or `reference\u002F` subtree** of any\n  kind. A mock or incomplete artifact in this skill's tree, even\n  one labeled \"reference\", is misleading: users will read it as\n  buildable.\n\n## Loading order\n\n1. Read this `SKILL.md` first to confirm the user's question is in\n   scope.\n2. **For the SHA capability matrix, algorithm enums, one-shot vs\n   partial task split, capability-query rules, permission matrix,\n   error taxonomy, observability, and safety \u002F path-selection\n   policy, see [CAPABILITIES.md](CAPABILITIES.md).**\n3. **For step-by-step workflows — configure, build, modify, run,\n   test, debug — see [TASKS.md](TASKS.md).**\n\nBoth companion files cross-link to each other,\n[`doca-version`](..\u002F..\u002Fdoca-version\u002FSKILL.md) for the canonical\nversion-handling rules, and\n[`doca-public-knowledge-map`](..\u002F..\u002Fdoca-public-knowledge-map\u002FSKILL.md)\nwhenever the right answer is \"look it up in the public docs or the\ninstalled package layout\" rather than \"SHA-specific guidance\".\n\n## Related skills\n\n- [`doca-public-knowledge-map`](..\u002F..\u002Fdoca-public-knowledge-map\u002FSKILL.md) —\n  the routing table for every public DOCA documentation source and\n  the on-disk layout of an installed DOCA package. The DOCA SHA\n  page lives at `docs.nvidia.com\u002Fdoca\u002Fsdk\u002FDOCA-SHA\u002F`; the File\n  Integrity reference application is the canonical worked example.\n- [`doca-setup`](..\u002F..\u002Fdoca-setup\u002FSKILL.md) — env preparation,\n  install verification, and the *I have no install yet* path with\n  the public NGC DOCA container. This skill assumes its\n  preconditions are satisfied.\n- [`doca-version`](..\u002F..\u002Fdoca-version\u002FSKILL.md) — canonical DOCA\n  version-handling rules. This skill's\n  [`## Version compatibility`](CAPABILITIES.md#version-compatibility)\n  cross-links the four-way match rule and adds only the SHA-specific\n  *\"discover algorithms + buffer sizes via cap query\"* overlay.\n- [`doca-structured-tools-contract`](..\u002F..\u002Fdoca-structured-tools-contract\u002FSKILL.md) —\n  the bundle's structured-tools precedence rule (detect \u002F prefer\n  \u002F fall back \u002F report). The Command appendix in\n  [TASKS.md](TASKS.md) honors this contract.\n- [`doca-programming-guide`](..\u002F..\u002Fdoca-programming-guide\u002FSKILL.md) —\n  general DOCA programming patterns shared by every library: the\n  canonical `pkg-config` + meson build pattern, the universal\n  modify-a-shipped-sample first-app workflow, the universal\n  lifecycle, the cross-library `DOCA_ERROR_*` taxonomy, and the\n  program-side debug order. This skill layers SHA specifics on\n  top.\n- [`doca-debug`](..\u002F..\u002Fdoca-debug\u002FSKILL.md) — the cross-cutting\n  debug ladder (install \u002F version \u002F build \u002F link \u002F runtime \u002F\n  program \u002F driver). SHA-specific debug (algorithm-not-supported,\n  destination-buffer-too-small, partial-hash-out-of-order) overlays\n  on top of that ladder.\n",{"data":35,"body":39},{"license":26,"name":4,"description":6,"metadata":36,"compatibility":38},{"kind":37},"library","Requires DOCA SDK installed at \u002Fopt\u002Fmellanox\u002Fdoca on Linux (Ubuntu 22.04\u002F24.04 or RHEL\u002FSLES) with a BlueField DPU or ConnectX NIC attached. Reads the user's local install via `pkg-config doca-sha` and inspects \u002Fopt\u002Fmellanox\u002Fdoca\u002F{lib,include,samples,applications}.\n",{"type":40,"children":41},"root",[42,50,139,146,158,529,535,562,602,608,613,805,827,833,845,934,948,954,966,1112,1118,1161,1186,1192],{"type":43,"tag":44,"props":45,"children":46},"element","h1",{"id":4},[47],{"type":48,"value":49},"text","DOCA SHA",{"type":43,"tag":51,"props":52,"children":53},"p",{},[54,60,62,67,69,80,82,88,90,99,101,106,108,118,120,125,127,137],{"type":43,"tag":55,"props":56,"children":57},"strong",{},[58],{"type":48,"value":59},"Where to start:",{"type":48,"value":61}," This skill assumes DOCA is already installed and\nthe user is doing ",{"type":43,"tag":55,"props":63,"children":64},{},[65],{"type":48,"value":66},"hands-on SHA-acceleration work",{"type":48,"value":68}," on a BlueField\n\u002F ConnectX \u002F host with DOCA. Open ",{"type":43,"tag":70,"props":71,"children":73},"a",{"href":72},"TASKS.md",[74],{"type":43,"tag":75,"props":76,"children":78},"code",{"className":77},[],[79],{"type":48,"value":72},{"type":48,"value":81}," if the user\nwants to ",{"type":43,"tag":83,"props":84,"children":85},"em",{},[86],{"type":48,"value":87},"do",{"type":48,"value":89}," something (configure \u002F build \u002F modify \u002F run \u002F test \u002F\ndebug); open ",{"type":43,"tag":70,"props":91,"children":93},{"href":92},"CAPABILITIES.md",[94],{"type":43,"tag":75,"props":95,"children":97},{"className":96},[],[98],{"type":48,"value":92},{"type":48,"value":100}," when the question\nis ",{"type":43,"tag":83,"props":102,"children":103},{},[104],{"type":48,"value":105},"what can DOCA SHA express",{"type":48,"value":107}," on this version. If the user has not\ninstalled DOCA yet, route to\n",{"type":43,"tag":70,"props":109,"children":111},{"href":110},"..\u002F..\u002Fdoca-setup\u002FSKILL.md",[112],{"type":43,"tag":75,"props":113,"children":115},{"className":114},[],[116],{"type":48,"value":117},"doca-setup",{"type":48,"value":119}," first. If the user is\nasking ",{"type":43,"tag":83,"props":121,"children":122},{},[123],{"type":48,"value":124},"\"should I even use the accelerator for this hash?\"",{"type":48,"value":126},", the\npath-selection rule in\n",{"type":43,"tag":70,"props":128,"children":130},{"href":129},"CAPABILITIES.md#capabilities-and-modes",[131],{"type":43,"tag":75,"props":132,"children":134},{"className":133},[],[135],{"type":48,"value":136},"CAPABILITIES.md ## Capabilities and modes",{"type":48,"value":138},"\nis the first stop.",{"type":43,"tag":140,"props":141,"children":143},"h2",{"id":142},"example-questions-this-skill-answers-well",[144],{"type":48,"value":145},"Example questions this skill answers well",{"type":43,"tag":51,"props":147,"children":148},{},[149,151,156],{"type":48,"value":150},"The CLASSES of DOCA SHA questions this skill is built to answer,\neach with one worked example. The agent should treat the ",{"type":43,"tag":83,"props":152,"children":153},{},[154],{"type":48,"value":155},"class",{"type":48,"value":157}," as\nthe load-bearing piece — the worked example is a single instance.",{"type":43,"tag":159,"props":160,"children":161},"ul",{},[162,215,291,334,394,450],{"type":43,"tag":163,"props":164,"children":165},"li",{},[166,171,173,178,180,188],{"type":43,"tag":55,"props":167,"children":168},{},[169],{"type":48,"value":170},"\"Should I offload this hash to DOCA SHA, or just compute it on\nthe CPU?\"",{"type":48,"value":172}," — worked example: ",{"type":43,"tag":83,"props":174,"children":175},{},[176],{"type":48,"value":177},"\"I am verifying file integrity on\na 4 GiB image; is doca-sha worth the setup vs OpenSSL on the\nCPU?\"",{"type":48,"value":179},". Answered by the path-selection table in\n",{"type":43,"tag":70,"props":181,"children":182},{"href":129},[183],{"type":43,"tag":75,"props":184,"children":186},{"className":185},[],[187],{"type":48,"value":136},{"type":43,"tag":159,"props":189,"children":190},{},[191],{"type":43,"tag":163,"props":192,"children":193},{},[194,196,201,203,213],{"type":48,"value":195},"the ",{"type":43,"tag":83,"props":197,"children":198},{},[199],{"type":48,"value":200},"\"when NOT to use doca-sha\"",{"type":48,"value":202}," bullets in\n",{"type":43,"tag":70,"props":204,"children":206},{"href":205},"CAPABILITIES.md#safety-policy",[207],{"type":43,"tag":75,"props":208,"children":210},{"className":209},[],[211],{"type":48,"value":212},"CAPABILITIES.md ## Safety policy",{"type":48,"value":214},".",{"type":43,"tag":163,"props":216,"children":217},{},[218,223,225,230,232,238,240,246,248,254,256,262,264,272],{"type":43,"tag":55,"props":219,"children":220},{},[221],{"type":48,"value":222},"\"Does my device support the SHA algorithm I want?\"",{"type":48,"value":224}," — worked\nexample: ",{"type":43,"tag":83,"props":226,"children":227},{},[228],{"type":48,"value":229},"\"is SHA-256 in the accelerator on this BlueField, and\nwhat is the minimum destination buffer size for it?\"",{"type":48,"value":231},". Answered\nby the algorithm + buffer-sizing capability-query rule\n(",{"type":43,"tag":75,"props":233,"children":235},{"className":234},[],[236],{"type":48,"value":237},"doca_sha_cap_task_hash_get_supported(devinfo, algorithm)",{"type":48,"value":239},"\nfor the one-shot path; ",{"type":43,"tag":75,"props":241,"children":243},{"className":242},[],[244],{"type":48,"value":245},"_task_partial_hash_get_supported(devinfo, algorithm)",{"type":48,"value":247},"\nfor the partial-hash path; ",{"type":43,"tag":75,"props":249,"children":251},{"className":250},[],[252],{"type":48,"value":253},"doca_sha_cap_get_min_dst_buf_size",{"type":48,"value":255},",\n",{"type":43,"tag":75,"props":257,"children":259},{"className":258},[],[260],{"type":48,"value":261},"doca_sha_cap_get_max_src_buf_size",{"type":48,"value":263},") in\n",{"type":43,"tag":70,"props":265,"children":266},{"href":129},[267],{"type":43,"tag":75,"props":268,"children":270},{"className":269},[],[271],{"type":48,"value":136},{"type":43,"tag":159,"props":273,"children":274},{},[275],{"type":43,"tag":163,"props":276,"children":277},{},[278,280,290],{"type":48,"value":279},"the discovery step in\n",{"type":43,"tag":70,"props":281,"children":283},{"href":282},"TASKS.md#configure",[284],{"type":43,"tag":75,"props":285,"children":287},{"className":286},[],[288],{"type":48,"value":289},"TASKS.md ## configure",{"type":48,"value":214},{"type":43,"tag":163,"props":292,"children":293},{},[294,299,300,305,307,315],{"type":43,"tag":55,"props":295,"children":296},{},[297],{"type":48,"value":298},"\"How do I pick between the one-shot and the partial \u002F incremental\nhash task?\"",{"type":48,"value":172},{"type":43,"tag":83,"props":301,"children":302},{},[303],{"type":48,"value":304},"\"my input is 1 GiB and the device\ncap says max source buffer is 64 MiB\"",{"type":48,"value":306},". Answered by the\none-shot-vs-partial table in\n",{"type":43,"tag":70,"props":308,"children":309},{"href":129},[310],{"type":43,"tag":75,"props":311,"children":313},{"className":312},[],[314],{"type":48,"value":136},{"type":43,"tag":159,"props":316,"children":317},{},[318],{"type":43,"tag":163,"props":319,"children":320},{},[321,323,333],{"type":48,"value":322},"the task-config workflow in\n",{"type":43,"tag":70,"props":324,"children":326},{"href":325},"TASKS.md#modify",[327],{"type":43,"tag":75,"props":328,"children":330},{"className":329},[],[331],{"type":48,"value":332},"TASKS.md ## modify",{"type":48,"value":214},{"type":43,"tag":163,"props":335,"children":336},{},[337,342,344,365,367,375],{"type":43,"tag":55,"props":338,"children":339},{},[340],{"type":48,"value":341},"\"What permissions does the source \u002F destination mmap need?\"",{"type":48,"value":343}," —\nworked example: ",{"type":43,"tag":83,"props":345,"children":346},{},[347,349,355,357,363],{"type":48,"value":348},"\"my ",{"type":43,"tag":75,"props":350,"children":352},{"className":351},[],[353],{"type":48,"value":354},"doca_sha_task_hash",{"type":48,"value":356}," returns\n",{"type":43,"tag":75,"props":358,"children":360},{"className":359},[],[361],{"type":48,"value":362},"DOCA_ERROR_NOT_PERMITTED",{"type":48,"value":364},"\"",{"type":48,"value":366},". Answered by the permission matrix\nin ",{"type":43,"tag":70,"props":368,"children":369},{"href":205},[370],{"type":43,"tag":75,"props":371,"children":373},{"className":372},[],[374],{"type":48,"value":212},{"type":43,"tag":159,"props":376,"children":377},{},[378],{"type":43,"tag":163,"props":379,"children":380},{},[381,383,393],{"type":48,"value":382},"the mmap-set-permissions checklist in\n",{"type":43,"tag":70,"props":384,"children":386},{"href":385},"TASKS.md#test",[387],{"type":43,"tag":75,"props":388,"children":390},{"className":389},[],[391],{"type":48,"value":392},"TASKS.md ## test",{"type":48,"value":214},{"type":43,"tag":163,"props":395,"children":396},{},[397,402,404,417,419,429,431,441,443,448],{"type":43,"tag":55,"props":398,"children":399},{},[400],{"type":48,"value":401},"\"Is this DOCA SHA API available on my installed DOCA version?\"",{"type":48,"value":403},"\n— worked example: ",{"type":43,"tag":83,"props":405,"children":406},{},[407,409,415],{"type":48,"value":408},"\"is ",{"type":43,"tag":75,"props":410,"children":412},{"className":411},[],[413],{"type":48,"value":414},"doca_sha_task_partial_hash",{"type":48,"value":416}," in the DOCA\nI have installed?\"",{"type":48,"value":418},". Answered by the version-compatibility\noverlay in\n",{"type":43,"tag":70,"props":420,"children":422},{"href":421},"CAPABILITIES.md#version-compatibility",[423],{"type":43,"tag":75,"props":424,"children":426},{"className":425},[],[427],{"type":48,"value":428},"CAPABILITIES.md ## Version compatibility",{"type":48,"value":430},",\nwhich cross-links the canonical detection chain in\n",{"type":43,"tag":70,"props":432,"children":434},{"href":433},"..\u002F..\u002Fdoca-version\u002FSKILL.md",[435],{"type":43,"tag":75,"props":436,"children":438},{"className":437},[],[439],{"type":48,"value":440},"doca-version",{"type":48,"value":442}," and adds the\nSHA-specific ",{"type":43,"tag":83,"props":444,"children":445},{},[446],{"type":48,"value":447},"\"discover, do not assume\"",{"type":48,"value":449}," bullets.",{"type":43,"tag":163,"props":451,"children":452},{},[453,466,467,486,488,498],{"type":43,"tag":55,"props":454,"children":455},{},[456,458,464],{"type":48,"value":457},"\"What does this ",{"type":43,"tag":75,"props":459,"children":461},{"className":460},[],[462],{"type":48,"value":463},"DOCA_ERROR_*",{"type":48,"value":465}," from a SHA call mean and which\nlayer caused it?\"",{"type":48,"value":172},{"type":43,"tag":83,"props":468,"children":469},{},[470,471,477,479,485],{"type":48,"value":364},{"type":43,"tag":75,"props":472,"children":474},{"className":473},[],[475],{"type":48,"value":476},"DOCA_ERROR_INVALID_VALUE",{"type":48,"value":478},"\non ",{"type":43,"tag":75,"props":480,"children":482},{"className":481},[],[483],{"type":48,"value":484},"doca_sha_task_hash_alloc_init",{"type":48,"value":364},{"type":48,"value":487},". Answered by the SHA overlay\non the cross-library taxonomy in\n",{"type":43,"tag":70,"props":489,"children":491},{"href":490},"CAPABILITIES.md#error-taxonomy",[492],{"type":43,"tag":75,"props":493,"children":495},{"className":494},[],[496],{"type":48,"value":497},"CAPABILITIES.md ## Error taxonomy",{"type":43,"tag":159,"props":499,"children":500},{},[501],{"type":43,"tag":163,"props":502,"children":503},{},[504,506,516,518,528],{"type":48,"value":505},"the layered ladder in\n",{"type":43,"tag":70,"props":507,"children":509},{"href":508},"TASKS.md#debug",[510],{"type":43,"tag":75,"props":511,"children":513},{"className":512},[],[514],{"type":48,"value":515},"TASKS.md ## debug",{"type":48,"value":517}," that escalates to\n",{"type":43,"tag":70,"props":519,"children":521},{"href":520},"..\u002F..\u002Fdoca-debug\u002FSKILL.md",[522],{"type":43,"tag":75,"props":523,"children":525},{"className":524},[],[526],{"type":48,"value":527},"doca-debug",{"type":48,"value":214},{"type":43,"tag":140,"props":530,"children":532},{"id":531},"audience",[533],{"type":48,"value":534},"Audience",{"type":43,"tag":51,"props":536,"children":537},{},[538,540,545,547,553,555,560],{"type":48,"value":539},"This skill serves ",{"type":43,"tag":55,"props":541,"children":542},{},[543],{"type":48,"value":544},"external developers building applications that\nconsume the DOCA SHA library",{"type":48,"value":546}," — i.e., users whose code calls\n",{"type":43,"tag":75,"props":548,"children":550},{"className":549},[],[551],{"type":48,"value":552},"doca_sha_*",{"type":48,"value":554}," (directly in C\u002FC++, or through FFI\u002Fbindings from\nanother language) to offload SHA hashing onto a BlueField DPU or\nConnectX accelerator. It is ",{"type":43,"tag":83,"props":556,"children":557},{},[558],{"type":48,"value":559},"not",{"type":48,"value":561}," for NVIDIA developers contributing\nto DOCA SHA itself.",{"type":43,"tag":51,"props":563,"children":564},{},[565,570,572,578,580,585,587,592,594,600],{"type":43,"tag":55,"props":566,"children":567},{},[568],{"type":48,"value":569},"Language scope.",{"type":48,"value":571}," DOCA SHA ships as a C library with ",{"type":43,"tag":75,"props":573,"children":575},{"className":574},[],[576],{"type":48,"value":577},"pkg-config",{"type":48,"value":579},"\nmodule name ",{"type":43,"tag":75,"props":581,"children":583},{"className":582},[],[584],{"type":48,"value":4},{"type":48,"value":586},". The shipped samples are written in C. C and\nC++ consumers are the canonical case and the worked examples in\n",{"type":43,"tag":75,"props":588,"children":590},{"className":589},[],[591],{"type":48,"value":72},{"type":48,"value":593}," assume that path. Other-language consumers (Rust, Go,\nPython, …) consume the same ",{"type":43,"tag":75,"props":595,"children":597},{"className":596},[],[598],{"type":48,"value":599},"*.so",{"type":48,"value":601}," through FFI or language-specific\nbindings; the skill's contribution in that case is to keep the\nlifecycle, capability-discovery, permission, error-taxonomy, and\none-shot-vs-partial guidance language-neutral, and to route the\nagent to the public C ABI as the authoritative surface that any\nwrapper will eventually call.",{"type":43,"tag":140,"props":603,"children":605},{"id":604},"when-to-load-this-skill",[606],{"type":48,"value":607},"When to load this skill",{"type":43,"tag":51,"props":609,"children":610},{},[611],{"type":48,"value":612},"Load this skill when the user is doing hands-on DOCA SHA work, in\nany language. Concretely:",{"type":43,"tag":159,"props":614,"children":615},{},[616,658,691,720,740,783,788,800],{"type":43,"tag":163,"props":617,"children":618},{},[619,621,627,629,635,637,642,644,649,651,657],{"type":48,"value":620},"Initializing a ",{"type":43,"tag":75,"props":622,"children":624},{"className":623},[],[625],{"type":48,"value":626},"doca_sha",{"type":48,"value":628}," context on a ",{"type":43,"tag":75,"props":630,"children":632},{"className":631},[],[633],{"type":48,"value":634},"doca_dev",{"type":48,"value":636}," and configuring\nat least one task type (",{"type":43,"tag":75,"props":638,"children":640},{"className":639},[],[641],{"type":48,"value":354},{"type":48,"value":643}," and\u002For\n",{"type":43,"tag":75,"props":645,"children":647},{"className":646},[],[648],{"type":48,"value":414},{"type":48,"value":650},") before ",{"type":43,"tag":75,"props":652,"children":654},{"className":653},[],[655],{"type":48,"value":656},"doca_ctx_start()",{"type":48,"value":214},{"type":43,"tag":163,"props":659,"children":660},{},[661,663,668,670,675,677,682,684,689],{"type":48,"value":662},"Choosing between the ",{"type":43,"tag":55,"props":664,"children":665},{},[666],{"type":48,"value":667},"one-shot",{"type":48,"value":669}," task (",{"type":43,"tag":75,"props":671,"children":673},{"className":672},[],[674],{"type":48,"value":354},{"type":48,"value":676}," —\ninput fits in a single source buffer, output digest lands in a\nsingle destination buffer) and the ",{"type":43,"tag":55,"props":678,"children":679},{},[680],{"type":48,"value":681},"partial \u002F incremental",{"type":48,"value":683}," task\n(",{"type":43,"tag":75,"props":685,"children":687},{"className":686},[],[688],{"type":48,"value":414},{"type":48,"value":690}," — input streamed in chunks, finalized\nseparately) for the user's data shape.",{"type":43,"tag":163,"props":692,"children":693},{},[694,696,702,704,710,712,718],{"type":48,"value":695},"Setting permissions on ",{"type":43,"tag":75,"props":697,"children":699},{"className":698},[],[700],{"type":48,"value":701},"doca_mmap",{"type":48,"value":703}," correctly for the source buffer\n(",{"type":43,"tag":75,"props":705,"children":707},{"className":706},[],[708],{"type":48,"value":709},"DOCA_ACCESS_FLAG_LOCAL_READ_ONLY",{"type":48,"value":711}," at minimum) and the destination\nbuffer (",{"type":43,"tag":75,"props":713,"children":715},{"className":714},[],[716],{"type":48,"value":717},"DOCA_ACCESS_FLAG_LOCAL_READ_WRITE",{"type":48,"value":719},").",{"type":43,"tag":163,"props":721,"children":722},{},[723,725,731,733,739],{"type":48,"value":724},"Sizing the destination buffer against\n",{"type":43,"tag":75,"props":726,"children":728},{"className":727},[],[729],{"type":48,"value":730},"doca_sha_cap_get_min_dst_buf_size(devinfo, algorithm)",{"type":48,"value":732}," and\nthe source buffer against\n",{"type":43,"tag":75,"props":734,"children":736},{"className":735},[],[737],{"type":48,"value":738},"doca_sha_cap_get_max_src_buf_size(devinfo)",{"type":48,"value":214},{"type":43,"tag":163,"props":741,"children":742},{},[743,745,751,753,759,760,766,768,773,775,781],{"type":48,"value":744},"Checking which SHA algorithm enums\n(",{"type":43,"tag":75,"props":746,"children":748},{"className":747},[],[749],{"type":48,"value":750},"DOCA_SHA_ALGORITHM_SHA1",{"type":48,"value":752},", ",{"type":43,"tag":75,"props":754,"children":756},{"className":755},[],[757],{"type":48,"value":758},"DOCA_SHA_ALGORITHM_SHA256",{"type":48,"value":255},{"type":43,"tag":75,"props":761,"children":763},{"className":762},[],[764],{"type":48,"value":765},"DOCA_SHA_ALGORITHM_SHA512",{"type":48,"value":767},") the active device's accelerator\nadvertises, via\n",{"type":43,"tag":75,"props":769,"children":771},{"className":770},[],[772],{"type":48,"value":237},{"type":48,"value":774}," and\n",{"type":43,"tag":75,"props":776,"children":778},{"className":777},[],[779],{"type":48,"value":780},"doca_sha_cap_task_partial_hash_get_supported(devinfo, algorithm)",{"type":48,"value":782}," —\nboth fold task-support and algorithm-support into one call.",{"type":43,"tag":163,"props":784,"children":785},{},[786],{"type":48,"value":787},"Validating a digest against a published test vector before pushing\nbulk input through the accelerator.",{"type":43,"tag":163,"props":789,"children":790},{},[791,793,798],{"type":48,"value":792},"Debugging a ",{"type":43,"tag":75,"props":794,"children":796},{"className":795},[],[797],{"type":48,"value":463},{"type":48,"value":799}," returned from a SHA call (lifecycle vs.\nbuffer-sizing vs. permission vs. unsupported-algorithm) and the\ntask-completion event on the progress engine.",{"type":43,"tag":163,"props":801,"children":802},{},[803],{"type":48,"value":804},"Designing or extending non-C bindings (Rust, Go, Python, …) that\nwrap the SHA C ABI — for the lifecycle, permission, capability,\nand one-shot-vs-partial rules the wrapper must honor.",{"type":43,"tag":51,"props":806,"children":807},{},[808,810,814,816,826],{"type":48,"value":809},"Do ",{"type":43,"tag":55,"props":811,"children":812},{},[813],{"type":48,"value":559},{"type":48,"value":815}," load this skill for general DOCA orientation, install of\nDOCA itself, non-SHA hashing libraries on CPU (use OpenSSL or\nsimilar), or other DOCA libraries. For those, use\n",{"type":43,"tag":70,"props":817,"children":819},{"href":818},"..\u002F..\u002Fdoca-public-knowledge-map\u002FSKILL.md",[820],{"type":43,"tag":75,"props":821,"children":823},{"className":822},[],[824],{"type":48,"value":825},"doca-public-knowledge-map",{"type":48,"value":214},{"type":43,"tag":140,"props":828,"children":830},{"id":829},"what-this-skill-provides",[831],{"type":48,"value":832},"What this skill provides",{"type":43,"tag":51,"props":834,"children":835},{},[836,838,843],{"type":48,"value":837},"This is a ",{"type":43,"tag":55,"props":839,"children":840},{},[841],{"type":48,"value":842},"thin loader",{"type":48,"value":844},". The body keeps only the orientation\nneeded to pick the right next file. The substantive SHA-specific\nmaterial lives in two companion files:",{"type":43,"tag":159,"props":846,"children":847},{},[848,873],{"type":43,"tag":163,"props":849,"children":850},{},[851,856,858,864,866,871],{"type":43,"tag":75,"props":852,"children":854},{"className":853},[],[855],{"type":48,"value":92},{"type":48,"value":857}," — what DOCA SHA can express on this version:\nthe two task types (one-shot hash and partial \u002F incremental\nhash), the three algorithm enums, the capability-query surface\n(",{"type":43,"tag":75,"props":859,"children":861},{"className":860},[],[862],{"type":48,"value":863},"doca_sha_cap_*",{"type":48,"value":865}," for algorithm support and buffer sizing), the\nSHA error taxonomy (mapped onto the cross-library ",{"type":43,"tag":75,"props":867,"children":869},{"className":868},[],[870],{"type":48,"value":463},{"type":48,"value":872},"\nset), the observability surface (per-task completion events on\nthe progress engine), the safety policy that gates source \u002F\ndestination mmap permission decisions, and the path-selection\nrule (when to use doca-sha versus a CPU hash or a different\nDOCA crypto library).",{"type":43,"tag":163,"props":874,"children":875},{},[876,881,883,889,890,896,897,903,904,910,911,917,918,924,926,932],{"type":43,"tag":75,"props":877,"children":879},{"className":878},[],[880],{"type":48,"value":72},{"type":48,"value":882}," — step-by-step workflows for the six in-scope SHA\nverbs: ",{"type":43,"tag":75,"props":884,"children":886},{"className":885},[],[887],{"type":48,"value":888},"configure",{"type":48,"value":752},{"type":43,"tag":75,"props":891,"children":893},{"className":892},[],[894],{"type":48,"value":895},"build",{"type":48,"value":752},{"type":43,"tag":75,"props":898,"children":900},{"className":899},[],[901],{"type":48,"value":902},"modify",{"type":48,"value":752},{"type":43,"tag":75,"props":905,"children":907},{"className":906},[],[908],{"type":48,"value":909},"run",{"type":48,"value":752},{"type":43,"tag":75,"props":912,"children":914},{"className":913},[],[915],{"type":48,"value":916},"test",{"type":48,"value":752},{"type":43,"tag":75,"props":919,"children":921},{"className":920},[],[922],{"type":48,"value":923},"debug",{"type":48,"value":925},".\nPlus a ",{"type":43,"tag":75,"props":927,"children":929},{"className":928},[],[930],{"type":48,"value":931},"Deferred task verbs",{"type":48,"value":933}," block that points out-of-scope\nquestions at the right next skill.",{"type":43,"tag":51,"props":935,"children":936},{},[937,939,947],{"type":48,"value":938},"The skill assumes a host or BlueField where DOCA is already\ninstalled at the standard location and the user has the privileges\ntheir public install profile expects. It does not cover installing\nDOCA — that path goes through\n",{"type":43,"tag":70,"props":940,"children":941},{"href":110},[942],{"type":43,"tag":75,"props":943,"children":945},{"className":944},[],[946],{"type":48,"value":117},{"type":48,"value":214},{"type":43,"tag":140,"props":949,"children":951},{"id":950},"what-this-skill-deliberately-does-not-ship",[952],{"type":48,"value":953},"What this skill deliberately does not ship",{"type":43,"tag":51,"props":955,"children":956},{},[957,959,964],{"type":48,"value":958},"This skill is ",{"type":43,"tag":55,"props":960,"children":961},{},[962],{"type":48,"value":963},"agent guidance",{"type":48,"value":965},", not a samples or templates\nbundle. To keep the boundary clean, it deliberately does not\ncontain — and pull requests should not add:",{"type":43,"tag":159,"props":967,"children":968},{},[969,1008,1032,1079],{"type":43,"tag":163,"props":970,"children":971},{},[972,977,979,985,987,997,999,1007],{"type":43,"tag":55,"props":973,"children":974},{},[975],{"type":48,"value":976},"Pre-written DOCA SHA application source code, in any\nlanguage.",{"type":48,"value":978}," The verified SHA source code is the shipped C samples\nat ",{"type":43,"tag":75,"props":980,"children":982},{"className":981},[],[983],{"type":48,"value":984},"\u002Fopt\u002Fmellanox\u002Fdoca\u002Fsamples\u002Fdoca_sha\u002F",{"type":48,"value":986},", plus the File Integrity\nreference application linked from the public DOCA SHA guide. The\nagent's job is to route the user to those files and prescribe a\nminimum-diff modification on them via the universal\nmodify-a-sample workflow in\n",{"type":43,"tag":70,"props":988,"children":990},{"href":989},"..\u002F..\u002Fdoca-programming-guide\u002FSKILL.md",[991],{"type":43,"tag":75,"props":992,"children":994},{"className":993},[],[995],{"type":48,"value":996},"doca-programming-guide",{"type":48,"value":998},",\nlayered with the SHA-specific overrides in\n",{"type":43,"tag":70,"props":1000,"children":1001},{"href":325},[1002],{"type":43,"tag":75,"props":1003,"children":1005},{"className":1004},[],[1006],{"type":48,"value":332},{"type":48,"value":214},{"type":43,"tag":163,"props":1009,"children":1010},{},[1011,1016,1018,1023,1025,1030],{"type":43,"tag":55,"props":1012,"children":1013},{},[1014],{"type":48,"value":1015},"Pre-computed digest tables for arbitrary inputs.",{"type":48,"value":1017}," The skill\ntells the agent to use a ",{"type":43,"tag":83,"props":1019,"children":1020},{},[1021],{"type":48,"value":1022},"published",{"type":48,"value":1024}," test vector (e.g. the NIST\nSHA test vectors for the empty string, \"abc\", and the\nmillion-",{"type":43,"tag":75,"props":1026,"children":1028},{"className":1027},[],[1029],{"type":48,"value":70},{"type":48,"value":1031}," input) as the known-vector smoke; it does not ship a\nvector bank of its own.",{"type":43,"tag":163,"props":1033,"children":1034},{},[1035,1040,1042,1048,1049,1055,1056,1062,1064,1069,1071,1077],{"type":43,"tag":55,"props":1036,"children":1037},{},[1038],{"type":48,"value":1039},"Standalone build manifests",{"type":48,"value":1041}," (",{"type":43,"tag":75,"props":1043,"children":1045},{"className":1044},[],[1046],{"type":48,"value":1047},"meson.build",{"type":48,"value":752},{"type":43,"tag":75,"props":1050,"children":1052},{"className":1051},[],[1053],{"type":48,"value":1054},"CMakeLists.txt",{"type":48,"value":255},{"type":43,"tag":75,"props":1057,"children":1059},{"className":1058},[],[1060],{"type":48,"value":1061},"Cargo.toml",{"type":48,"value":1063},", …) parked inside the skill. The agent constructs\nthe build manifest ",{"type":43,"tag":83,"props":1065,"children":1066},{},[1067],{"type":48,"value":1068},"in the user's project directory",{"type":48,"value":1070}," against the\nuser's installed DOCA, where ",{"type":43,"tag":75,"props":1072,"children":1074},{"className":1073},[],[1075],{"type":48,"value":1076},"pkg-config --modversion doca-sha",{"type":48,"value":1078},"\nis the source of truth.",{"type":43,"tag":163,"props":1080,"children":1081},{},[1082,1110],{"type":43,"tag":55,"props":1083,"children":1084},{},[1085,1087,1093,1094,1100,1102,1108],{"type":48,"value":1086},"A ",{"type":43,"tag":75,"props":1088,"children":1090},{"className":1089},[],[1091],{"type":48,"value":1092},"samples\u002F",{"type":48,"value":752},{"type":43,"tag":75,"props":1095,"children":1097},{"className":1096},[],[1098],{"type":48,"value":1099},"bindings\u002F",{"type":48,"value":1101},", or ",{"type":43,"tag":75,"props":1103,"children":1105},{"className":1104},[],[1106],{"type":48,"value":1107},"reference\u002F",{"type":48,"value":1109}," subtree",{"type":48,"value":1111}," of any\nkind. A mock or incomplete artifact in this skill's tree, even\none labeled \"reference\", is misleading: users will read it as\nbuildable.",{"type":43,"tag":140,"props":1113,"children":1115},{"id":1114},"loading-order",[1116],{"type":48,"value":1117},"Loading order",{"type":43,"tag":1119,"props":1120,"children":1121},"ol",{},[1122,1135,1148],{"type":43,"tag":163,"props":1123,"children":1124},{},[1125,1127,1133],{"type":48,"value":1126},"Read this ",{"type":43,"tag":75,"props":1128,"children":1130},{"className":1129},[],[1131],{"type":48,"value":1132},"SKILL.md",{"type":48,"value":1134}," first to confirm the user's question is in\nscope.",{"type":43,"tag":163,"props":1136,"children":1137},{},[1138],{"type":43,"tag":55,"props":1139,"children":1140},{},[1141,1143,1147],{"type":48,"value":1142},"For the SHA capability matrix, algorithm enums, one-shot vs\npartial task split, capability-query rules, permission matrix,\nerror taxonomy, observability, and safety \u002F path-selection\npolicy, see ",{"type":43,"tag":70,"props":1144,"children":1145},{"href":92},[1146],{"type":48,"value":92},{"type":48,"value":214},{"type":43,"tag":163,"props":1149,"children":1150},{},[1151],{"type":43,"tag":55,"props":1152,"children":1153},{},[1154,1156,1160],{"type":48,"value":1155},"For step-by-step workflows — configure, build, modify, run,\ntest, debug — see ",{"type":43,"tag":70,"props":1157,"children":1158},{"href":72},[1159],{"type":48,"value":72},{"type":48,"value":214},{"type":43,"tag":51,"props":1162,"children":1163},{},[1164,1166,1174,1176,1184],{"type":48,"value":1165},"Both companion files cross-link to each other,\n",{"type":43,"tag":70,"props":1167,"children":1168},{"href":433},[1169],{"type":43,"tag":75,"props":1170,"children":1172},{"className":1171},[],[1173],{"type":48,"value":440},{"type":48,"value":1175}," for the canonical\nversion-handling rules, and\n",{"type":43,"tag":70,"props":1177,"children":1178},{"href":818},[1179],{"type":43,"tag":75,"props":1180,"children":1182},{"className":1181},[],[1183],{"type":48,"value":825},{"type":48,"value":1185},"\nwhenever the right answer is \"look it up in the public docs or the\ninstalled package layout\" rather than \"SHA-specific guidance\".",{"type":43,"tag":140,"props":1187,"children":1189},{"id":1188},"related-skills",[1190],{"type":48,"value":1191},"Related skills",{"type":43,"tag":159,"props":1193,"children":1194},{},[1195,1216,1236,1267,1288,1315],{"type":43,"tag":163,"props":1196,"children":1197},{},[1198,1206,1208,1214],{"type":43,"tag":70,"props":1199,"children":1200},{"href":818},[1201],{"type":43,"tag":75,"props":1202,"children":1204},{"className":1203},[],[1205],{"type":48,"value":825},{"type":48,"value":1207}," —\nthe routing table for every public DOCA documentation source and\nthe on-disk layout of an installed DOCA package. The DOCA SHA\npage lives at ",{"type":43,"tag":75,"props":1209,"children":1211},{"className":1210},[],[1212],{"type":48,"value":1213},"docs.nvidia.com\u002Fdoca\u002Fsdk\u002FDOCA-SHA\u002F",{"type":48,"value":1215},"; the File\nIntegrity reference application is the canonical worked example.",{"type":43,"tag":163,"props":1217,"children":1218},{},[1219,1227,1229,1234],{"type":43,"tag":70,"props":1220,"children":1221},{"href":110},[1222],{"type":43,"tag":75,"props":1223,"children":1225},{"className":1224},[],[1226],{"type":48,"value":117},{"type":48,"value":1228}," — env preparation,\ninstall verification, and the ",{"type":43,"tag":83,"props":1230,"children":1231},{},[1232],{"type":48,"value":1233},"I have no install yet",{"type":48,"value":1235}," path with\nthe public NGC DOCA container. This skill assumes its\npreconditions are satisfied.",{"type":43,"tag":163,"props":1237,"children":1238},{},[1239,1247,1249,1258,1260,1265],{"type":43,"tag":70,"props":1240,"children":1241},{"href":433},[1242],{"type":43,"tag":75,"props":1243,"children":1245},{"className":1244},[],[1246],{"type":48,"value":440},{"type":48,"value":1248}," — canonical DOCA\nversion-handling rules. This skill's\n",{"type":43,"tag":70,"props":1250,"children":1251},{"href":421},[1252],{"type":43,"tag":75,"props":1253,"children":1255},{"className":1254},[],[1256],{"type":48,"value":1257},"## Version compatibility",{"type":48,"value":1259},"\ncross-links the four-way match rule and adds only the SHA-specific\n",{"type":43,"tag":83,"props":1261,"children":1262},{},[1263],{"type":48,"value":1264},"\"discover algorithms + buffer sizes via cap query\"",{"type":48,"value":1266}," overlay.",{"type":43,"tag":163,"props":1268,"children":1269},{},[1270,1280,1282,1286],{"type":43,"tag":70,"props":1271,"children":1273},{"href":1272},"..\u002F..\u002Fdoca-structured-tools-contract\u002FSKILL.md",[1274],{"type":43,"tag":75,"props":1275,"children":1277},{"className":1276},[],[1278],{"type":48,"value":1279},"doca-structured-tools-contract",{"type":48,"value":1281}," —\nthe bundle's structured-tools precedence rule (detect \u002F prefer\n\u002F fall back \u002F report). The Command appendix in\n",{"type":43,"tag":70,"props":1283,"children":1284},{"href":72},[1285],{"type":48,"value":72},{"type":48,"value":1287}," honors this contract.",{"type":43,"tag":163,"props":1289,"children":1290},{},[1291,1299,1301,1306,1308,1313],{"type":43,"tag":70,"props":1292,"children":1293},{"href":989},[1294],{"type":43,"tag":75,"props":1295,"children":1297},{"className":1296},[],[1298],{"type":48,"value":996},{"type":48,"value":1300}," —\ngeneral DOCA programming patterns shared by every library: the\ncanonical ",{"type":43,"tag":75,"props":1302,"children":1304},{"className":1303},[],[1305],{"type":48,"value":577},{"type":48,"value":1307}," + meson build pattern, the universal\nmodify-a-shipped-sample first-app workflow, the universal\nlifecycle, the cross-library ",{"type":43,"tag":75,"props":1309,"children":1311},{"className":1310},[],[1312],{"type":48,"value":463},{"type":48,"value":1314}," taxonomy, and the\nprogram-side debug order. This skill layers SHA specifics on\ntop.",{"type":43,"tag":163,"props":1316,"children":1317},{},[1318,1326],{"type":43,"tag":70,"props":1319,"children":1320},{"href":520},[1321],{"type":43,"tag":75,"props":1322,"children":1324},{"className":1323},[],[1325],{"type":48,"value":527},{"type":48,"value":1327}," — the cross-cutting\ndebug ladder (install \u002F version \u002F build \u002F link \u002F runtime \u002F\nprogram \u002F driver). SHA-specific debug (algorithm-not-supported,\ndestination-buffer-too-small, partial-hash-out-of-order) overlays\non top of that ladder.",{"items":1329,"total":1433},[1330,1347,1361,1375,1387,1404,1419],{"slug":1331,"name":1331,"fn":1332,"description":1333,"org":1334,"tags":1335,"stars":23,"repoUrl":24,"updatedAt":1346},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1336,1339,1342,1343],{"name":1337,"slug":1338,"type":15},"Data Analysis","data-analysis",{"name":1340,"slug":1341,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":1344,"slug":1345,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":1348,"name":1348,"fn":1349,"description":1350,"org":1351,"tags":1352,"stars":23,"repoUrl":24,"updatedAt":1360},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1353,1356,1359],{"name":1354,"slug":1355,"type":15},"Deployment","deployment",{"name":1357,"slug":1358,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1362,"name":1362,"fn":1363,"description":1364,"org":1365,"tags":1366,"stars":23,"repoUrl":24,"updatedAt":1374},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1367,1370,1371],{"name":1368,"slug":1369,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1372,"slug":1373,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1376,"name":1376,"fn":1377,"description":1378,"org":1379,"tags":1380,"stars":23,"repoUrl":24,"updatedAt":1386},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1381,1382,1383],{"name":1337,"slug":1338,"type":15},{"name":9,"slug":8,"type":15},{"name":1384,"slug":1385,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":1388,"name":1388,"fn":1389,"description":1390,"org":1391,"tags":1392,"stars":23,"repoUrl":24,"updatedAt":1403},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1393,1396,1399,1400],{"name":1394,"slug":1395,"type":15},"Automation","automation",{"name":1397,"slug":1398,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1401,"slug":1402,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1405,"name":1405,"fn":1406,"description":1407,"org":1408,"tags":1409,"stars":23,"repoUrl":24,"updatedAt":1418},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1410,1411,1414,1415],{"name":1354,"slug":1355,"type":15},{"name":1412,"slug":1413,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1416,"slug":1417,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1420,"name":1420,"fn":1421,"description":1422,"org":1423,"tags":1424,"stars":23,"repoUrl":24,"updatedAt":1432},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1425,1426,1429],{"name":9,"slug":8,"type":15},{"name":1427,"slug":1428,"type":15},"Quantum Computing","quantum-computing",{"name":1430,"slug":1431,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":1435,"total":1586},[1436,1454,1470,1481,1493,1507,1520,1534,1545,1554,1568,1577],{"slug":1437,"name":1437,"fn":1438,"description":1439,"org":1440,"tags":1441,"stars":1451,"repoUrl":1452,"updatedAt":1453},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1442,1445,1448],{"name":1443,"slug":1444,"type":15},"Documentation","documentation",{"name":1446,"slug":1447,"type":15},"MCP","mcp",{"name":1449,"slug":1450,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1455,"name":1455,"fn":1456,"description":1457,"org":1458,"tags":1459,"stars":1467,"repoUrl":1468,"updatedAt":1469},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1460,1463,1464],{"name":1461,"slug":1462,"type":15},"Containers","containers",{"name":1354,"slug":1355,"type":15},{"name":1465,"slug":1466,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1471,"name":1471,"fn":1472,"description":1473,"org":1474,"tags":1475,"stars":1467,"repoUrl":1468,"updatedAt":1480},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1476,1479],{"name":1477,"slug":1478,"type":15},"CI\u002FCD","ci-cd",{"name":1354,"slug":1355,"type":15},"2026-07-14T05:25:59.97109",{"slug":1482,"name":1482,"fn":1483,"description":1484,"org":1485,"tags":1486,"stars":1467,"repoUrl":1468,"updatedAt":1492},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1487,1488,1489],{"name":1477,"slug":1478,"type":15},{"name":1354,"slug":1355,"type":15},{"name":1490,"slug":1491,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1494,"name":1494,"fn":1495,"description":1496,"org":1497,"tags":1498,"stars":1467,"repoUrl":1468,"updatedAt":1506},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1499,1502,1503],{"name":1500,"slug":1501,"type":15},"Debugging","debugging",{"name":1490,"slug":1491,"type":15},{"name":1504,"slug":1505,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1508,"name":1508,"fn":1509,"description":1510,"org":1511,"tags":1512,"stars":1467,"repoUrl":1468,"updatedAt":1519},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1513,1516],{"name":1514,"slug":1515,"type":15},"Best Practices","best-practices",{"name":1517,"slug":1518,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1521,"name":1521,"fn":1522,"description":1523,"org":1524,"tags":1525,"stars":1467,"repoUrl":1468,"updatedAt":1533},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1526,1529,1532],{"name":1527,"slug":1528,"type":15},"Machine Learning","machine-learning",{"name":1530,"slug":1531,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1535,"name":1535,"fn":1536,"description":1537,"org":1538,"tags":1539,"stars":1467,"repoUrl":1468,"updatedAt":1544},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1540,1543],{"name":1541,"slug":1542,"type":15},"QA","qa",{"name":1384,"slug":1385,"type":15},"2026-07-14T05:25:53.673039",{"slug":1546,"name":1546,"fn":1547,"description":1548,"org":1549,"tags":1550,"stars":1467,"repoUrl":1468,"updatedAt":1553},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1551,1552],{"name":1354,"slug":1355,"type":15},{"name":1357,"slug":1358,"type":15},"2026-07-14T05:25:49.362534",{"slug":1555,"name":1555,"fn":1556,"description":1557,"org":1558,"tags":1559,"stars":1467,"repoUrl":1468,"updatedAt":1567},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1560,1563,1564],{"name":1561,"slug":1562,"type":15},"Code Review","code-review",{"name":1490,"slug":1491,"type":15},{"name":1565,"slug":1566,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1569,"name":1569,"fn":1570,"description":1571,"org":1572,"tags":1573,"stars":1467,"repoUrl":1468,"updatedAt":1576},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1574,1575],{"name":1541,"slug":1542,"type":15},{"name":1384,"slug":1385,"type":15},"2026-07-14T05:25:54.928983",{"slug":1578,"name":1578,"fn":1579,"description":1580,"org":1581,"tags":1582,"stars":1467,"repoUrl":1468,"updatedAt":1585},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1583,1584],{"name":1394,"slug":1395,"type":15},{"name":1477,"slug":1478,"type":15},"2026-07-30T05:29:03.275638",496]