New: 97,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Hyperloom Workload Optimizer

by amd in amd/skills

Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer.

MITAuto-check: notesAI & LLM Engineering

Install Hyperloom Workload Optimizer

skills CLI
$ npx skills add amd/skills --skill hyperloom-workload-optimizer -a claude-code

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

GitHub CLI
$ gh skill install amd/skills hyperloom-workload-optimizer --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/amd/skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/hyperloom-workload-optimizer .claude/skills/hyperloom-workload-optimizer && 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/

Facts

Skill name
hyperloom-workload-optimizer
GitHub stars
395
Token cost
~1.7k tokens
SKILL.md length
815 words
Files
5
Skills in repo
9
Repo updated
First seen
Licence
MIT

At a glance

Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer.

  • Works in 4 steps: Confirm the workspace → Install the Hyperloom wheel → Run /hyperloom-setup → …
  • The user wants to make a model serve faster
  • SKILL.md covers Out of scope for this skill, Prerequisites, Step 1: Confirm the workspace and Step 2: Install the Hyperloom…, plus 4 more sections
  • Calls pip and python

What it does

Hyperloom Workload Optimizer is an agent skill from amd/skills. Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer. Given a model, framework, workload (TP/EP, concurrency, ISL/OSL, precision), an objective and a time budget, it explores per-workload which levers to pull (serving/config parameters and env, framework enablement and source patches, and hot GPU-kernel rewrites), benchmarks each candidate, and returns the optimization stack that produced the gain. Use when the user…

Its SKILL.md is about 1.7k tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files (for example `.federated.json`, `evals/evals.json` and `evals/machine.yml`).

It sits in AI & LLM Engineering, covering LLM inference and serving and Meeting notes and agendas. It works with SGLang, vLLM and Python. The repository describes itself as: Official AMD catalog of AI agent skills. Empower your AI agents with AMD's optimized SW stack. The licence is MIT.

When your agent uses it

  • The user wants to make a model serve faster
  • Raise tokens/sec
  • SGLang on MI300X/MI325X/MI355X
  • Run the kernel-agent

Example prompts

  • “/hyperloom-workload-optimizer”

Requirements

  • Python 3
  • Docker

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Confirm the workspace
  2. Install the Hyperloom wheel
  3. Run /hyperloom-setup
  4. Hand off to a run skill

What it can do on your machine

Read from SKILL.md and the folder at commit 6c92b41. 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:

    • pip
    • python

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use pip, which can reach the network depending on how they are called.

    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

Hyperloom Workload Optimizer loads about 1.7k tokens when it runs. Until then it costs about 243 tokens; SKILL.md has 815 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~243
When it runs · the whole SKILL.md, loaded when a task matches
~1.7k

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: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:61
    updates `.env` in it. Do not switch to another directory on your own, and do not
  • NoteMentions a .env fileSKILL.md:62
    ing project unless the user accepts the `.env` change.
  • NoteMentions a .env fileSKILL.md:85
    framework install. It writes `.env` and stops before any optimization. Run it once
  • NoteMentions a .env fileSKILL.md:110
    session, follow the optimizer skill at `.env`

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 amd/skills at commit 6c92b41, republished under its MIT licence (© amd). 815 words, ~1,689 tokens.

Download SKILL.mdSave it as .claude/skills/hyperloom-workload-optimizer/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
hyperloom-workload-optimizer
description
Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer. Given a model, framework, workload (TP/EP, concurrency, ISL/OSL, precision), an objective and a time budget, it explores per-workload which levers to pull (serving/config parameters and env, framework enablement and source patches, and hot GPU-kernel rewrites), benchmarks each candidate, and returns the optimization stack that produced the gain. Use when the user wants to make a model serve faster, raise tokens/sec or throughput, optimize or tune vLLM or SGLang on MI300X/MI325X/MI355X, run Hyperloom, run the kernel-agent, quantize-then-optimize with Quark, set up Hyperloom from scratch, or resume a Hyperloom session. Do not use to stand up a server for plain serving, diagnose a broken ROCm install, or run a one-off kernel/benchmark or trace analysis without the optimization loop.
<!--
Copyright (c) 2026 Advanced Micro Devices, Inc. All rights reserved.

See LICENSE for license information.
-->

Hyperloom Workload Optimizer

You are the entry point for Hyperloom optimization on AMD Instinct GPUs. Your job is the bootstrap: confirm the workspace, install the Hyperloom wheel, run /hyperloom-setup, then hand the run to the skill that owns it.

The wheel installs the skills that own everything after setup: hyperloom-setup for credentials and run mode, the demo skills for a workload preset, and inference_optimizer for the launcher gates and monitoring. They ship with the runtime, so they always match the installed version.

Out of scope for this skill

  • Do not run python -m hyperloom.inference_optimizer.cli optimize yourself.
  • Do not implement a GPU preflight, launcher gate, or background launch here. The installed skills own those, including the Iron Rules and the resume path.
  • Do not ask for workload values (model, TP/EP, concurrency, ISL/OSL, precision, objective, budget) while installing or while setup is running. They belong to the run skill, after setup finishes.
  • Do not optimize by hand in chat.

Prerequisites

  • AMD Instinct GPU host (MI300X / MI325X / MI355X) with ROCm, /dev/kfd and /dev/dri present, and amd-smi or rocm-smi working.
  • Python 3.10+ and pip on the machine that runs the install.
  • LLM credentials: Anthropic API access, or the AMD LLM gateway.
  • A dedicated empty directory, opened in the agent as the workspace.

Confirm the shell you are in is on the GPU host before installing. Setup may later point Docker at a different target host; until it does, everything here runs where the agent is.

Step 1: Confirm the workspace

The current directory is both the install target and the agent workspace. Confirm with the user that it is a dedicated directory before installing: setup creates or updates .env in it. Do not switch to another directory on your own, and do not install into an existing project unless the user accepts the .env change.

Step 2: Install the Hyperloom wheel

bash
pip install hyperloom-inference-optimizer --target .

Install the current release unless the user asks for a specific version. It is normal for the directory to hold many Python package folders afterwards; the user does not need to inspect them.

Step 3: Run /hyperloom-setup

The wheel installs hyperloom-setup into the agent's skill directories (.agents/skills/, .claude/skills/, .cursor/skills/). Run it:

text
/hyperloom-setup

It is interactive and owns credentials, USER_DATA_PATH, the run mode (docker recommended, or baremetal), the Docker target host, and the bare-metal framework install. It writes .env and stops before any optimization. Run it once per workspace; the run skills reuse those values.

Let setup ask its own questions. Do not preempt them, do not restate its option lists, and never ask the user to paste an API key into chat.

If the agent does not list hyperloom-setup after the install, the skill directories were written after the agent scanned them. Tell the user to restart the agent, then run it again. Do not substitute your own setup steps.

Show full SKILL.md (349 more words)Show less

Step 4: Hand off to a run skill

Setup ends by offering a run and loading the matching skill, so normally you just follow it. When the user asks for a run directly, load the skill by name and follow its instructions instead of this one:

  • hyperloom-qwen3-8b-3h — short no-kernel Qwen3-8B run; best first end-to-end check.
  • hyperloom-qwen3-14b-fp8-12h — medium-length Qwen3-14B-FP8 run.
  • hyperloom-qwen3-14b-fp8-12h-forge — the same run on the KernelForge kernel backend.
  • hyperloom-custom-advanced — explicit model, framework, workload, budget, and phase toggles.

A preset keeps its workload even if the user supplies their own MODEL_PATH; tensor parallelism, concurrency, sequence lengths, precision, and budget are not retuned for that model. When those need to change, use hyperloom-custom-advanced.

To resume a stopped session, follow the optimizer skill at .env HYPERLOOM_SKILL_PATH; it owns the resume path and the gates a relaunch still has to clear.

What to expect during a run

Optimization runs for hours in the background. Do not stream the log.

Before launch the run skill shows a plan: resolved model path, run mode, framework, TP, concurrency, ISL/OSL, precision, budget, and USER_DATA_PATH. Get the user's go-ahead on that plan before the optimizer starts — it then owns the GPU for hours. After launch it reports the optimizer PID, run log, launch-info JSON, session directory, state.json, and the first health check.

During the run, report a short status about every 300 seconds: process alive, current phase, stop_reason, baseline and current best throughput, cumulative gain, the latest benchmark or candidate decision, and the most relevant log lines. Never print API keys, tokens, or custom headers.

Troubleshooting

  • Many package folders in the workspace after pip install --target . is expected.
  • /hyperloom-setup not listed: the install landed after the agent scanned for skills. Restart the agent and check .claude/skills/hyperloom-setup/ exists.
  • ImportError: libamdhip64.so.7 or libhipblas.so.3: the framework torch wheel wants different ROCm user-space libraries; align ROCM_PATH and LD_LIBRARY_PATH.
  • hipDeviceAttributePciChipId missing during an AITER build: hipcc is using older ROCm headers; put the matching ROCm bin first on PATH.
  • Anything past setup (preflight failures, launch, phases, gains) belongs to the installed run and optimizer skills. Read those rather than reproducing their checks here.

© amd, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 4 other files in skills/hyperloom-workload-optimizer of amd/skills.

  • SKILL.md
  • .federated.json
  • evals/evals.json
  • evals/machine.yml
  • skill-card.md

Open the folder on GitHubat commit 6c92b41

Compare with similar skills

Hyperloom Workload Optimizer 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.

Hyperloom Workload Optimizer compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Hyperloom Workload Optimizer this skillamd/skills395—~1.7kAutomated safety check: NotesMIT
Dstack Prototypingdstackai/dstack2.3k—~1.6kAutomated safety check: PassMPL-2.0
One EvalOpenDCAI/One-Eval165—~2.4kAutomated safety check: PassApache-2.0
SGLang Structured ServingOrchestra-Research/AI-Research-SKILLs13k3 repos~2.9kAutomated safety check: PassMIT
LLM Serving Framework BenchmarkBBuf/AI-Infra-Auto-Driven-SKILLS900—~7.5kAutomated safety check: PassNone
Hyperloom SetupAMD-AGI/Hyperloom216—~7.2kAutomated safety check: NotesCustom licence

Similar skills

  • Dstack Prototyping

    dstackai/dstack

    Use with the dstack skill for model-serving work when the image, serving command, resources, backend/fleet choice, or service behavior is not proven.

    2.3k GitHub stars~1.6k tokensUpdated today
    AI & LLM EngineeringAuto-check passed

  • One Eval

    OpenDCAI/One-Eval

    驱动 One-Eval 对 API 或本地模型做端到端评测，覆盖纯文本、多模态、代码生成、函数调用和 Agent benchmark。当用户想评测模型在一个或多个 benchmark 上的表现、比较分数、补充 metric，或生成图文评测报告时使用本 skill。

    165 GitHub stars~2.4k tokensUpdated 1 mo ago
    AI & LLM EngineeringAuto-check passed

  • SGLang Structured Serving

    Orchestra-Research/AI-Research-SKILLs

    Covers serving LLMs with SGLang, whose RadixAttention reuses cached prefixes, and constraining output to JSON, regex or grammar for agent and tool-calling workloads.

    13k GitHub starsUsed in 3 repos~2.9k tokens
    AI & LLM EngineeringAuto-check passed

  • LLM Serving Framework Benchmark

    BBuf/AI-Infra-Auto-Driven-SKILLS

    Compares SGLang, vLLM, TensorRT-LLM and TokenSpeed on one model and workload, searching server flags to find the best deployment command within a latency SLA.

    900 GitHub stars~7.5k tokensUpdated 2 days ago
    AI & LLM EngineeringAuto-check passed

  • Hyperloom Setup

    AMD-AGI/Hyperloom

    Configures Hyperloom after pip install --target . An agent skill from AMD-AGI/Hyperloom.

    216 GitHub stars~7.2k tokensUpdated today
    DevOps & CloudAuto-check: notes

  • SageMaker Serving Image Selection

    huggingface/skills

    Official

    Chooses the right serving container and current image URI for deploying a Hugging Face model to a SageMaker endpoint, preferring Hugging Face images over generic ones.

    11k GitHub starsUsed in 1 repo~4.6k tokens
    AI & LLM EngineeringAuto-check passed

More from amd/skills

All 9 skills in this repo

  • Apu Memory Tuner

    amd/skills

    Inspects and tunes the shared-vs-dedicated memory split on AMD Ryzen APUs with unified memory (UMA) so larger LLMs and image-gen models fit on the iGPU, or so reserved GPU memory is returned to the…

    395 GitHub stars~2.6k tokensUpdated today
    Auto-check passed

  • Lemonade Router Builder

    amd/skills

    Turns a natural-language description of routing intent into a valid Lemonade collection.router policy JSON.

    395 GitHub stars~4k tokensUpdated today
    Auto-check passed

  • Local AI Use

    amd/skills

    Makes this agent generate images, transcribe audio, and synthesize speech on the user's own machine through a local Lemonade Server instead of a paid cloud API.

    395 GitHub stars~5k tokensUpdated today
    Auto-check: notes

  • Serving LLMs On Instinct

    amd/skills

    Serves AI models on AMD Instinct GPU hardware using vLLM. An agent skill from amd/skills.

    395 GitHub stars~4k tokensUpdated today
    Auto-check: notes

  • Serving LLMs On Epyc

    amd/skills

    Serves an LLM on a supported AMD EPYC server CPU using vLLM with zentorch, in Docker, Podman, or conda.

    395 GitHub stars~5.7k tokensUpdated today
    Auto-check: notes

  • Magpie Kernel Evaluator

    amd/skills

    Benchmarks LLM inference and drives GPU kernel optimization with Magpie.

    395 GitHub stars~2.3k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Hyperloom Workload Optimizer

What does Hyperloom Workload Optimizer do?

Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer. Hyperloom Workload Optimizer is an agent skill from amd/skills. Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer.

When should I use Hyperloom Workload Optimizer?

Hyperloom Workload Optimizer fits situations like: the user wants to make a model serve faster; raise tokens/sec; SGLang on MI300X/MI325X/MI355X; run the kernel-agent.

How do I install Hyperloom Workload Optimizer in Claude Code?

Run `npx skills add amd/skills --skill hyperloom-workload-optimizer -a claude-code`. Or copy the skill folder (skills/hyperloom-workload-optimizer in amd/skills) into .claude/skills/hyperloom-workload-optimizer in your project. Claude Code loads it when a task matches its description.

How do I install Hyperloom Workload Optimizer in Codex?

Run `npx skills add amd/skills --skill hyperloom-workload-optimizer -a codex`. Or copy the skill folder (skills/hyperloom-workload-optimizer in amd/skills) into .agents/skills/hyperloom-workload-optimizer in your project. Codex loads it when a task matches its description.

Can I use Hyperloom Workload Optimizer 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 amd/skills --skill hyperloom-workload-optimizer -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/hyperloom-workload-optimizer, .gemini/skills/hyperloom-workload-optimizer, .github/skills/hyperloom-workload-optimizer and .opencode/skills/hyperloom-workload-optimizer in your project.

What does Hyperloom Workload Optimizer need to run?

Going by SKILL.md and its folder, Hyperloom Workload Optimizer needs the command-line tools its instructions call (pip and python). Our summary lists: Python 3; Docker.

Does Hyperloom Workload Optimizer access the network?

SKILL.md contains no URLs. Its commands use pip, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Hyperloom Workload Optimizer safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Hyperloom Workload Optimizer use?

Hyperloom Workload Optimizer is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Hyperloom Workload Optimizer use?

About 1.7k tokens (SKILL.md is roughly 6.8k 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 Hyperloom Workload Optimizer?

Skills that share tags, products or a category with Hyperloom Workload Optimizer: Dstack Prototyping (dstackai/dstack, 2.3k stars), One Eval (OpenDCAI/One-Eval, 165 stars), SGLang Structured Serving (Orchestra-Research/AI-Research-SKILLs, 13k stars) and LLM Serving Framework Benchmark (BBuf/AI-Infra-Auto-Driven-SKILLS, 900 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Hyperloom Workload Optimizer?

amd (a GitHub organization) maintains it in amd/skills, which has 395 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on October 7, 2026.

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