New: 98,000+ 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

Hf Gguf Quant Jobs

by Mesh-LLM in Mesh-LLM/mesh-llm

A skill your agent uses when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos…

Apache-2.0Auto-check passedAI & LLM Engineering

Install Hf Gguf Quant Jobs

skills CLI
$ npx skills add Mesh-LLM/mesh-llm --skill hf-gguf-quant-jobs -a claude-code

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

GitHub CLI
$ gh skill install Mesh-LLM/mesh-llm hf-gguf-quant-jobs --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/Mesh-LLM/mesh-llm.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/hf-gguf-quant-jobs .claude/skills/hf-gguf-quant-jobs && 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
hf-gguf-quant-jobs
GitHub stars
3.5k
Token cost
~1.9k tokens
SKILL.md length
761 words
Files
2
Skills in repo
25
Repo updated
First seen
Licence
Apache-2.0

At a glance

A skill your agent uses when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos…

  • Works in 8 steps: Identify the source BF16/FP16 GGUF repo,… → Preflight both Hub and mounted source… → Write or upload a quant-plan.json with… → …
  • Documenting low-memory Hugging Face Jobs
  • SKILL.md covers Preconditions, Workflow, Launch Template and Monitoring, plus 2 more sections
  • Calls hf and just; needs HF_TOKEN

What it does

Hf Gguf Quant Jobs is an agent skill from Mesh-LLM/mesh-llm. Use when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos with skippy-quantize.

Its SKILL.md is about 1.9k 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 Model hubs and datasets. It works with llama.cpp and Hugging Face. The repository describes itself as: Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat. The licence is Apache-2.0.

When your agent uses it

  • Documenting low-memory Hugging Face Jobs
  • Local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos with skippy-quantize

Example prompts

  • “/hf-gguf-quant-jobs”

Workflow steps

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

  1. Identify the source BF16/FP16 GGUF repo, target quant repo, output prefix,
  2. Preflight both Hub and mounted source paths with skippy-quantize status,
  3. Write or upload a quant-plan.json with source repo/revision, target repo,
  4. Launch the job with --window-size 1 for the first full model run unless a
  5. For each split window, stage only the required input shard, run
  6. Monitor for progress markers. A healthy job repeatedly emits staged source
  7. Validate the target repo after completion by counting GGUF shards, checking
  8. Record the artifact in the experiment card and create an iteration card for

What it can do on your machine

Read from SKILL.md and the folder at commit 1b9f0cf. 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:

    • hf
    • just

    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 these keys or tokens, usually read from environment variables:

    • HF_TOKEN

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Hf Gguf Quant Jobs loads about 1.9k tokens when it runs. Until then it costs about 54 tokens; SKILL.md has 761 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~54
When it runs · the whole SKILL.md, loaded when a task matches
~1.9k

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 Mesh-LLM/mesh-llm at commit 1b9f0cf, republished under its Apache-2.0 licence (© Mesh-LLM). 761 words, ~1,890 tokens.

Download SKILL.mdSave it as .claude/skills/hf-gguf-quant-jobs/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
hf-gguf-quant-jobs
description
Use when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos with skippy-quantize.

HF GGUF Quant Jobs

Use this skill to turn an existing split BF16/FP16 GGUF model repo into a quantized GGUF model repo without requiring the host to hold the full model in memory or on local disk at once. The operational tool is skippy-quantize; do not use llama-quantize, llama-quantise, or wrapper scripts that shell out to those binaries.

The supported pattern is: mount or point at the source BF16/FP16 GGUF repo, quantize resumable split windows with skippy-quantize, publish completed output shards to the target model repo, delete staged files immediately, and resume from the first missing target shard after cancellation or failure.

Preconditions

  • Use a split BF16/FP16 GGUF repo as the source when possible. Do not re-read SafeTensors for requants if a BF16 GGUF artifact already exists.
  • Verify the source repo is complete before spending on quantization. Count all expected split shards and refuse to run if any are missing.
  • Use a tensor-type file for any custom recipe. Treat MTP tensors, output tensors, precision-sensitive tensors, and latency-sensitive layer ranges as explicit recipe inputs.
  • Run jobs under the intended HF org and pass HF_TOKEN as a secret, not a printed environment variable.
  • Prefer mounted Hub repos over full hf download when the job only needs to stream or stage one shard/window at a time.
  • Build the standalone binary with just skippy-quantize-standalone-release-build for local runs or in the job image/script for HF Jobs.

Workflow

  1. Identify the source BF16/FP16 GGUF repo, target quant repo, output prefix, output basename, source prefix, quant type, tensor-type file, memory budget, and split window size.
  2. Preflight both Hub and mounted source paths with skippy-quantize status, next-window, validate-splits, or a quantize --preflight-only run. Stop if the source artifact is incomplete.
  3. Write or upload a quant-plan.json with source repo/revision, target repo, quant type, shard count, output prefix, tensor policy, and resume settings.
  4. Launch the job with --window-size 1 for the first full model run unless a smaller fixture proves a larger window is safe on the chosen hardware.
  5. For each split window, stage only the required input shard, run skippy-quantize run-quant-window or run-quant, publish finished shards, then delete local staged input and output files.
  6. Monitor for progress markers. A healthy job repeatedly emits staged source copies, quant_window, publish completion, cleanup, and increasing split progress.
  7. Validate the target repo after completion by counting GGUF shards, checking the first and last shard names, and confirming quant-plan.json plus the tensor-type file are present.
  8. Record the artifact in the experiment card and create an iteration card for the run, including job id, command, environment, repo SHA, shard count, and follow-up decisions.
Show full SKILL.md (326 more words)Show less

Launch Template

Create a quantization manifest:

bash
target/release/skippy-quantize init-quant \
  --source /mnt/source-gguf \
  --source-prefix <source-prefix> \
  --target /mnt/target-quant \
  --target-prefix <target-prefix> \
  --output-basename <output-basename> \
  --quant <quant> \
  --tensor-type-file /mnt/recipe/tensor-types.txt \
  --window-size 1 \
  --manifest /tmp/skippy-quantize.json

Dry-run the next quantization window before spending I/O:

bash
target/release/skippy-quantize quant-job \
  --source /mnt/source-gguf \
  --source-prefix <source-prefix> \
  --target /mnt/target-quant \
  --target-prefix <target-prefix> \
  --output-basename <output-basename> \
  --quant <quant> \
  --tensor-type-file /mnt/recipe/tensor-types.txt \
  --window-size 1 \
  --manifest /tmp/skippy-quantize.json \
  --backend skippy-abi \
  --max-memory 32G \
  --dry-run

Run until complete:

bash
target/release/skippy-quantize run-quant \
  --manifest /tmp/skippy-quantize.json \
  --backend skippy-abi \
  --max-memory 32G \
  --work-dir /tmp/skippy-quantize-work \
  --spool-dir /tmp/skippy-quantize-output \
  --record-dir /tmp/skippy-quantize-records \
  --json-event-file /tmp/skippy-quantize-status.json \
  --json-event-interval-seconds 120 \
  --json-event-window 8

For HF Jobs, mount the BF16/FP16 source repo and target quant repo, then run the same manifest and run-quant commands inside the job:

bash
hf jobs uv run \
  --namespace meshllm \
  --flavor cpu-upgrade \
  --timeout 3d \
  --secrets HF_TOKEN \
  --volume hf://models/<source-repo>:/mnt/source-gguf \
  --volume hf://models/<target-repo>:/mnt/target-quant \
  --env SKIPPY_QUANTIZE_OUTPUT=json \
  --env PYTHONUNBUFFERED=1 \
  --detach \
  /path/to/skippy_quant_job.py \
  -- \
  --source /mnt/source-gguf \
  --source-prefix <source-prefix> \
  --target /mnt/target-quant \
  --target-prefix <target-prefix> \
  --output-basename <output-basename> \
  --quant <quant> \
  --tensor-type-file /mnt/recipe/tensor-types.txt \
  --max-memory 32G

The job script should only build or install skippy-quantize, prepare the manifest if missing, run run-quant, verify the job, and upload sidecars.

Monitoring

Check status and logs:

bash
hf jobs inspect <job-id> --namespace meshllm
hf jobs logs <job-id> --namespace meshllm --tail 120

For agents, prefer polling /tmp/skippy-quantize-status.json over ingesting full logs. It is a periodically refreshed compact snapshot with the current phase, current split window, and a bounded recent-event window.

Useful healthy markers:

  • Preflight QuantizeGguf with backend skippy-abi
  • Source artifact is complete
  • quant_window
  • Published /mnt/target-quant/...
  • Cleaned staged source
  • split artifact ... 100.00%

Concerning markers:

  • repeated watchdog lines with no shard, tensor, upload, or cache-drop progress;
  • cgroup memory pinned near the hardware limit;
  • the same split window restarting repeatedly without new uploaded target files;
  • fallback quant warnings for tensors that the recipe expected to preserve.

If a job stalls, cancel it before changing code or hardware. The next run should skip already published shards and resume at the first missing output shard.

Validation

After completion, verify the target repo with an authenticated Hub API or CLI check. Record at least:

  • target repo and commit SHA;
  • privacy setting;
  • total file count;
  • GGUF shard count;
  • first and last shard names;
  • manifest/plan presence;
  • tensor-type file presence.

For local smoke tests, use a small split GGUF source first and verify:

  • skippy-quantize verify-job --manifest <manifest> --llama-load succeeds;
  • skippy-quantize validate-splits --root <target> --prefix <prefix> succeeds;
  • max RSS stays bounded compared with full-model size;
  • skippy-quantize status --manifest <manifest> --json reports completion.

Documentation Contract

For Jianyang-style experiments, update both records:

  • the main experiment card with the promoted artifact;
  • a phase iteration card with the job id, exact command, environment, verification output, decision, and follow-ups.

Keep post-experiment upstream notes separate from the run decision. The job can be successful while the converter or quantizer patches still need extraction into clean upstream PRs.

© Mesh-LLM, Apache-2.0. 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 1 other file in .agents/skills/hf-gguf-quant-jobs of Mesh-LLM/mesh-llm.

  • SKILL.md
  • agents/openai.yaml

Open the folder on GitHubat commit 1b9f0cf

Compare with similar skills

Hf Gguf Quant Jobs 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.

Hf Gguf Quant Jobs compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Hf Gguf Quant Jobs this skillMesh-LLM/mesh-llm3.5k—~1.9kAutomated safety check: PassApache-2.0
Qwen Mtp GgufR6410418/Jackrong-llm-finetuning-guide1.7k—~1.7kAutomated safety check: PassMIT
Hugging Face LLM Trainerhuggingface/skills11k1 repos~7.2kAutomated safety check: PassApache-2.0
Hugging Face Local Modelshuggingface/skills11k3 repos~945Automated safety check: PassApache-2.0
Add Modelguoqingbao/xinfer334—~4.2kAutomated safety check: NotesMIT
Huggingface LLM Trainerwaybarrios/opencode-power-pack534—~3kAutomated safety check: PassApache-2.0

Similar skills

  • Qwen Mtp Gguf

    R6410418/Jackrong-llm-finetuning-guide

    Complete agent-ready workflow for Qwen-family MTP or nextn GGUF conversion and release.

    1.7k GitHub stars~1.7k tokensUpdated 3 mo ago
    AI & LLM EngineeringAuto-check passed

  • Hugging Face LLM Trainer

    huggingface/skills

    Official

    Trains or fine-tunes language and vision models with TRL or Unsloth on Hugging Face Jobs cloud GPUs, then converts the results to GGUF.

    11k GitHub starsUsed in 1 repo~7.2k tokens
    AI & LLM EngineeringAuto-check passed

  • Hugging Face Local Models

    huggingface/skills

    Official

    Finds llama.cpp-compatible GGUF models on the Hugging Face Hub, picks a quantization for your hardware and launches them with llama-cli or llama-server.

    11k GitHub starsUsed in 3 repos~945 tokens
    AI & LLM EngineeringAuto-check passed

  • Add Model

    guoqingbao/xinfer

    Adapt and port new LLM model architectures to this xinfer project.

    334 GitHub stars~4.2k tokensUpdated 1 mo ago
    AI & LLM EngineeringAuto-check: notes

  • Huggingface LLM Trainer

    waybarrios/opencode-power-pack

    Train or fine-tune language models with TRL or Unsloth on Hugging Face Jobs, including SFT, DPO, GRPO, reward models, and GGUF conversion.

    534 GitHub stars~3k tokensUpdated 5 days ago
    AI & LLM EngineeringAuto-check passed

  • Resolve

    alexziskind1/model-shelf

    Always resolve Hugging Face models via model-shelf before any download.

    130 GitHub stars~792 tokensUpdated 1 mo ago
    AI & LLM EngineeringAuto-check passed

More from Mesh-LLM/mesh-llm

All 25 skills in this repo

  • Release Validation

    Mesh-LLM/mesh-llm

    A skill your agent uses when validating a MeshLLM release candidate or current HEAD against the last GitHub release, assembling the canonical feature/fix/modification inventory, testing locally…

    3.5k GitHub stars~2.6k tokensUpdated today
    Auto-check passed

  • Benchmark Tune

    Mesh-LLM/mesh-llm

    A skill your agent uses when running, debugging, interpreting, or documenting mesh-llm benchmark tune model-serving throughput trials, including choosing…

    3.5k GitHub stars~1.6k tokensUpdated today
    Auto-check passed

  • Config Settings Management

    Mesh-LLM/mesh-llm

    A skill your agent uses when adding, renaming, removing, validating, or exposing mesh-llm config settings, including built-in settings, plugin config schemas, owner-control apply behavior, CLI…

    3.5k GitHub stars~1.4k tokensUpdated today
    Auto-check passed

  • Connect Agents

    Mesh-LLM/mesh-llm

    A skill your agent uses when connecting agent tools or OpenAI clients to mesh-llm — launching or configuring Goose, Claude Code, OpenCode, Pi, curl, or any OpenAI-compatible client against a local…

    3.5k GitHub stars~1.2k tokensUpdated today
    Auto-check passed

  • Hf Bf16 Gguf Conversion Jobs

    Mesh-LLM/mesh-llm

    A skill your agent uses when converting Hugging Face SafeTensors checkpoints into split BF16 GGUF model repos with skippy-quantize on Hugging Face Jobs or a local machine, then publishing the…

    3.5k GitHub stars~1.1k tokensUpdated today
    Auto-check passed

  • Hf Layer Package Jobs

    Mesh-LLM/mesh-llm

    A skill your agent uses when changing mesh-llm automation or CLI flows that discover Hugging Face GGUF models, plan CPU Hugging Face Jobs for layer-package splitting, estimate max cost, or publish…

    3.5k GitHub stars~1.4k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Hf Gguf Quant Jobs

What does Hf Gguf Quant Jobs do?

A skill your agent uses when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos…. Hf Gguf Quant Jobs is an agent skill from Mesh-LLM/mesh-llm. Use when creating, monitoring, validating, or documenting low-memory Hugging Face Jobs or local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos with skippy-quantize.

When should I use Hf Gguf Quant Jobs?

Hf Gguf Quant Jobs fits situations like: documenting low-memory Hugging Face Jobs; local runs that quantize split BF16/FP16 GGUF model repos into custom quant GGUF repos with skippy-quantize.

How do I install Hf Gguf Quant Jobs in Claude Code?

Run `npx skills add Mesh-LLM/mesh-llm --skill hf-gguf-quant-jobs -a claude-code`. Or copy the skill folder (.agents/skills/hf-gguf-quant-jobs in Mesh-LLM/mesh-llm) into .claude/skills/hf-gguf-quant-jobs in your project. Claude Code loads it when a task matches its description.

How do I install Hf Gguf Quant Jobs in Codex?

Run `npx skills add Mesh-LLM/mesh-llm --skill hf-gguf-quant-jobs -a codex`. Or copy the skill folder (.agents/skills/hf-gguf-quant-jobs in Mesh-LLM/mesh-llm) into .agents/skills/hf-gguf-quant-jobs in your project. Codex loads it when a task matches its description.

Can I use Hf Gguf Quant Jobs 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 Mesh-LLM/mesh-llm --skill hf-gguf-quant-jobs -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/hf-gguf-quant-jobs, .gemini/skills/hf-gguf-quant-jobs, .github/skills/hf-gguf-quant-jobs and .opencode/skills/hf-gguf-quant-jobs in your project.

What does Hf Gguf Quant Jobs need to run?

Going by SKILL.md and its folder, Hf Gguf Quant Jobs needs the command-line tools its instructions call (hf and just) and credentials named HF_TOKEN.

Does Hf Gguf Quant Jobs 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 Hf Gguf Quant Jobs 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 Hf Gguf Quant Jobs use?

Hf Gguf Quant Jobs 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 Hf Gguf Quant Jobs use?

About 1.9k tokens (SKILL.md is roughly 7.6k 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 Hf Gguf Quant Jobs?

Skills that share tags, products or a category with Hf Gguf Quant Jobs: Qwen Mtp Gguf (R6410418/Jackrong-llm-finetuning-guide, 1.7k stars), Hugging Face LLM Trainer (huggingface/skills, 11k stars), Hugging Face Local Models (huggingface/skills, 11k stars) and Add Model (guoqingbao/xinfer, 334 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Hf Gguf Quant Jobs?

Mesh-LLM (a GitHub organization) maintains it in Mesh-LLM/mesh-llm, which has 3,495 GitHub stars. The repository holds 25 skills in this directory. The repository was last updated on October 11, 2026.

Source: Mesh-LLM/mesh-llm on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.