Agent skill

Llama Nnc

Develop and profile the edge-e3 PyTorch-to-NNC flow, including nnc/compiler.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and…

Apache-2.0Auto-check passedAI & LLM Engineering

Install Llama Nnc

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.codex/skills/llama-nnc .claude/skills/llama-nnc && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .claude/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.codex/skills/llama-nnc .agents/skills/llama-nnc && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .agents/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.codex/skills/llama-nnc .cursor/skills/llama-nnc && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .cursor/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/exeex/edge-cores.git --path .codex/skills/llama-nnc

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.codex/skills/llama-nnc .gemini/skills/llama-nnc && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .gemini/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .github/skills && cp -r skills-src/.codex/skills/llama-nnc .github/skills/llama-nnc && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .github/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add exeex/edge-cores --skill llama-nnc -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install exeex/edge-cores llama-nnc --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/exeex/edge-cores.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.codex/skills/llama-nnc .opencode/skills/llama-nnc && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "llama-nnc" agent skill from https://github.com/exeex/edge-cores/tree/main/.codex/skills/llama-nnc into .opencode/skills/llama-nnc/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "llama-nnc", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: llama-nnc
GitHub stars: 110
Token cost: ~1.1k tokens
SKILL.md length: 498 words
Files: 2
Skills in repo: 9
Repo updated: Sep 26, 2026
First seen: Oct 7, 2026
Licence: Apache-2.0

At a glance

Develop and profile the edge-e3 PyTorch-to-NNC flow, including nnc/compiler.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and…

Works in 5 steps: Define or update its PyTorch custom op… → Update target mapping in nnc/graph.py,… → Render its C++ call in nnc/codegen.py's… → …
Tasks that involve Deep learning
SKILL.md covers Read first, Development workflow, Correctness and memory rules and Transformer-block profiling…, plus 1 more section
Calls python3

What it does

Llama Nnc is an agent skill from exeex/edge-cores. Develop and profile the edge-e3 PyTorch-to-NNC flow, including nnc/compiler.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and per-node cycle reports on the encrypted Verilator core.

Its SKILL.md is about 1.1k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).

It sits in AI & LLM Engineering, covering Deep learning. It works with C++ and PyTorch. The repository describes itself as: The shortest path from PyTorch to custom ASICs. The licence is Apache-2.0.

When your agent uses it

Tasks that involve Deep learning

Example prompts

“/llama-nnc”

Requirements

Python 3

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

Define or update its PyTorch custom op in nnc/test/smoke_.py.
Update target mapping in nnc/graph.py, lowering metadata and shape
Render its C++ call in nnc/codegen.py's ForwardRenderer.
Implement it in cpp/libnn/.hpp and expose it from cpp/libnn/ops.hpp.
Run the compiler tests and the smallest relevant encrypted-core smoke.

What it can do on your machine

Read from SKILL.md and the folder at commit 68dc8aa. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
- python3
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Llama Nnc loads about 1.1k tokens when it runs. Until then it costs about 65 tokens; SKILL.md has 498 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~65
When it runs · the whole SKILL.md, loaded when a task matches: ~1.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from exeex/edge-cores at commit 68dc8aa, republished under its Apache-2.0 licence (© exeex). 498 words, ~1,128 tokens.

Download SKILL.mdSave it as .claude/skills/llama-nnc/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.

name: llama-nnc
description: Develop and profile the edge-e3 PyTorch-to-NNC flow, including nnc/compiler.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and per-node cycle reports on the encrypted Verilator core.

Llama NNC

Keep this repository's NNC path bare-metal-only and runnable with the public encrypted core.

Read first

Read example/llama/README.md for the user-facing flow and artifact locations.
Read nnc/compiler.py and the affected module before changing export (graph.py), lowering and DRAM preferences (lowering.py), liveness (liveness.py), weight packing (weights.py), or generated headers (codegen.py).
Read the affected header in cpp/libnn/ before changing an operator.
Read .codex/skills/edge-tensor-example/SKILL.md for Tensor/DMA intrinsic changes and .codex/skills/edge-verilator-demo/SKILL.md for simulator changes.

Do not add an example-local libnn, startup, host runner, NNEDGE_HOST branch, private RTL dependency, or old CMake harness. Use cpp/libnn/, cpp/baremetal/, and scripts/build-verilator.sh.

Development workflow

For a semantic operator change:

Define or update its PyTorch custom op in nnc/test/smoke_<op>.py.
Update target mapping in nnc/graph.py, lowering metadata and shape inference in nnc/lowering.py, or weight packing in nnc/weights.py only where needed.
Render its C++ call in nnc/codegen.py's ForwardRenderer.
Implement it in cpp/libnn/<op>.hpp and expose it from cpp/libnn/ops.hpp.
Run the compiler tests and the smallest relevant encrypted-core smoke.

python3 -m unittest nnc.test_compiler
./example/llama/run.sh --model-file nnc/test/smoke_<op>.py

Set PYTHON=/path/to/python for the shell wrapper, or invoke nnc/run_smoke.py with that interpreter directly. Install dependencies from nnc/requirements.txt.

Correctness and memory rules

Preserve PyTorch contiguous public input/output byte order. Treat packed or tiled accelerator buffers as operator-private layouts.
Pack only static Linear weights in the compiler. Keep Linear input/output tensors contiguous.
Pass values between operators only through Tensor objects. Never retain a dtcm_op_scratch() pointer as tensor storage.
Keep the first 96 KiB of DTCM for the tensor arena and the final 32 KiB for synchronous operator scratch. Check worst-case byte sizes.
Treat free_tensor() as LIFO-only reclamation and never use a cleared tensor.
Do not use DMA for DTCM-to-DTCM copies. nnedge::copy() uses CMPU COPY mode 9 for aligned BF16 DTCM transfers.
Clean DRAM sources before DMA, synchronize producers before consumers, and clean/invalidate DRAM destinations around accelerator output DMA.
Do not move tensor payloads with scalar loads/stores. Scalar code may manage metadata and initialize operator-owned scratch only.
Do not generalize shapes beyond layouts covered by composed Verilator tests.

Show full SKILL.md (172 more words)Show less

Transformer-block profiling smoke

Run example/llama/model/llama3_source.py directly as the complete tiny Llama transformer block with compiler-owned per-node instrumentation:

./example/llama/profile.sh

The compiler measures immediately around each lowered operator call, completes the final output copy, and prints all records afterward. Do not add temporary cycle reads or printf calls inside operators for node-level reports.

Require all of these results:

encrypted Verilator exits with TEST PASS;
all output BF16 values match the PyTorch golden within the runner tolerance;
profile.md and profile.tsv exist beside the ELF;
software-console.log retains the raw profile records.

Interpret each percentage as a share of the sum of measured operator calls, not total simulation cycles. Allocation, input/final copies, startup, and printing are excluded. Preserve node indices and names so repeated operations remain distinguishable.

Validation

After compiler or runtime changes, run:

python3 -m unittest nnc.test_compiler
./example/llama/run.sh
./example/llama/test.sh

The suite must cover every nnc/test/smoke_*.py file and finish by profiling nnc/test/llama3_source.py. Keep that file byte-identical to the user-facing example/llama/model/llama3_source.py. Require harness.md to report every row as PASS; do not maintain a silent skip list.

Generated files belong under example/llama/build/ and must remain untracked.

Files

SKILL.md and 1 other file in .codex/skills/llama-nnc of exeex/edge-cores.

SKILL.md
agents/openai.yaml

Open the folder on GitHubat commit 68dc8aa

Compare with similar skills

Llama Nnc next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Llama Nnc compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Llama Nnc this skillexeex/edge-cores	110	—	~1.1k	Automated safety check: Pass	Apache-2.0	13 days ago
Ako4allTongmingLAIC/AKO4ALL	369	—	~4k	Automated safety check: Pass	MIT	24 days ago
Paddle Op DevPaddlePaddle/Paddle	24k	—	~1.3k	Automated safety check: Pass	Apache-2.0	today
Qualcomm QNN Backend Developmentpytorch/executorch	5.1k	—	~1.8k	Automated safety check: Pass	Custom licence	today
Embedded AI Deploymentmatlab/agent-skills-playground	183	—	~3.4k	Automated safety check: Pass	Custom licence	today
Paddle Cross Ecosystem Custom OpPaddlePaddle/Paddle	24k	—	~883	Automated safety check: Pass	Apache-2.0	today

Similar skills

Ako4all
TongmingLAIC/AKO4ALL
Drive an agentic loop that iteratively optimizes a GPU kernel for maximum speedup.
369 GitHub stars~4k tokensUpdated 24 days ago
AI & LLM EngineeringAuto-check passed
Paddle Op Dev
PaddlePaddle/Paddle
PaddlePaddle (飞桨) C++ 算子开发指南。提供从 YAML 配置、InferMeta 函数、Kernel 实现、Python API 封装、单元测试到编译验证的完整算子开发流程指导。在以下场景使用此 skill：(1) 为 Paddle 框架新增 C++ 算子 (2) 修改或调试已有 Paddle 算子 (3) 编写算子的 YAML…
24k GitHub stars~1.3k tokensUpdated today
AI & LLM EngineeringAuto-check passed
Qualcomm QNN Backend Development
pytorch/executorch
Helps build, test and extend the Qualcomm AI Engine Direct (QNN) backend in ExecuTorch, with routes for new ops, model export, Buck-vs-CMake parity fixes and per-layer accuracy debugging.
5.1k GitHub stars~1.8k tokensUpdated today
AI & LLM EngineeringAuto-check passed
Embedded AI Deployment
matlab/agent-skills-playground
Deploy AI models to embedded hardware using MathWorks tools (MATLAB, Simulink, Embedded Coder).
183 GitHub stars~3.4k tokensUpdated today
AI & LLM EngineeringAuto-check passed
Paddle Cross Ecosystem Custom Op
PaddlePaddle/Paddle
将原生 PyTorch 自定义算子库、Torch extension、生态库（TorchCodec/FlashInfer/DeepEP 等）以及 Kernel DSL 生态（Triton/TileLang/TVM FFI 等）以最小修改方式接入 PaddlePaddle。遇到以下场景务必使用：迁移外部算子库到 Paddle；分析 PFCCLab fork 与上游的兼容差异；处理…
24k GitHub stars~883 tokensUpdated today
AI & LLM EngineeringAuto-check passed
At Dispatch V2
intel/torch-xpu-ops
Official
Convert PyTorch ATDISPATCH macros to ATDISPATCHV2 format in ATen C++ code.
115 GitHub starsUsed in 3 repos~2.2k tokens
AI & LLM EngineeringAuto-check passed

Questions about Llama Nnc

What does Llama Nnc do?

Develop and profile the edge-e3 PyTorch-to-NNC flow, including nnc/compiler.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and…. Llama Nnc is an agent skill from exeex/edge-cores.py lowering and generated ABI, example/llama smoke models, cpp/libnn runtime operators, BF16 correctness checks, and per-node cycle reports on the encrypted Verilator core.

When should I use Llama Nnc?

Llama Nnc fits situations like: tasks that involve Deep learning.

How do I install Llama Nnc in Claude Code?

Run `npx skills add exeex/edge-cores --skill llama-nnc -a claude-code`. Or copy the skill folder (.codex/skills/llama-nnc in exeex/edge-cores) into .claude/skills/llama-nnc in your project. Claude Code loads it when a task matches its description.

How do I install Llama Nnc in Codex?

Run `npx skills add exeex/edge-cores --skill llama-nnc -a codex`. Or copy the skill folder (.codex/skills/llama-nnc in exeex/edge-cores) into .agents/skills/llama-nnc in your project. Codex loads it when a task matches its description.

Can I use Llama Nnc in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add exeex/edge-cores --skill llama-nnc -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/llama-nnc, .gemini/skills/llama-nnc, .github/skills/llama-nnc and .opencode/skills/llama-nnc in your project.

What does Llama Nnc need to run?

Going by SKILL.md and its folder, Llama Nnc needs the command-line tools its instructions call (python3). Our summary lists: Python 3.

Does Llama Nnc access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Llama Nnc safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Llama Nnc use?

Llama Nnc is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Llama Nnc use?

About 1.1k tokens (SKILL.md is roughly 4.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Llama Nnc?

Skills that share tags, products or a category with Llama Nnc: Ako4all (TongmingLAIC/AKO4ALL, 369 stars), Paddle Op Dev (PaddlePaddle/Paddle, 24k stars), Qualcomm QNN Backend Development (pytorch/executorch, 5.1k stars) and Embedded AI Deployment (matlab/agent-skills-playground, 183 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Llama Nnc?

exeex (a GitHub user) maintains it in exeex/edge-cores, which has 110 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on September 26, 2026.

Source: exeex/edge-cores on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Llama Nnc

Install Llama Nnc

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

Llama NNC

Read first

Development workflow

Correctness and memory rules

Transformer-block profiling smoke

Validation

Files

Compare with similar skills

Similar skills

Ako4all

Paddle Op Dev

Qualcomm QNN Backend Development

Embedded AI Deployment

Paddle Cross Ecosystem Custom Op

At Dispatch V2

More from exeex/edge-cores

Edge Bringup

Edge Contribution

Rtl Release Packager

Integrate Your Design

Edge Synth Report

Software Harness

Questions about Llama Nnc

Install Llama Nnc

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

Llama NNC

Read first

Development workflow

Correctness and memory rules

Transformer-block profiling smoke

Validation

Files

Compare with similar skills

Similar skills

More from exeex/edge-cores

Related

Questions about Llama Nnc