Waxell

skill

unsloth High risk

orchestra-researchorchestra-research/ai-research-skillsupdated

Browse GitHub Download zipBook a demo
Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization
orchestra-research's own description, from the skill's files.
ResearchDocumentsAI & agentsskill
Stars
14,234
Version
v1
Copies
0
OverviewSecurity25Versions1

Overview

blockHigh risk

Waxell rates it high risk: 25 findings, highest severity critical.

Highest severity
critical
Findings
25
What ran
Static analysis

A clean result means nothing alarming in what is published, not a guarantee. How the checks work

We don't republish the skill's instructions. Read them on GitHub at the reviewed commit.

Security

Security scan

Checked 2 days ago

Result

High risk Waxell rates it high risk: 25 findings, highest severity critical.

Checked with static analysis: servers it launches, packages it installs at runtime, scripts piped to a shell and bundled binaries. Nothing was run. How the checks work

Findings (25)

critical · 1

  • critical

    Tells Claude to download a script and pipe it to a shell

    references/llms-full.md:1029 · skill_risk:S1

high · 21

  • high

    Tells Claude to install unpinned packages (unsloth)

    references/llms-full.md:62 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (unsloth, unsloth_zoo)

    references/llms-full.md:934 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (git+https://github.com/unslothai/unsloth.git)

    references/llms-full.md:939 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (git+https://github.com/unslothai/unsloth-zoo.git)

    references/llms-full.md:940 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (git+https://github.com/unslothai/unsloth_zoo.git)

    references/llms-full.md:967 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (pip)

    references/llms-full.md:1000 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (torch)

    references/llms-full.md:1036 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (torch, torchvision, torchaudio, triton)

    references/llms-full.md:1389 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (notebook)

    references/llms-full.md:1489 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (torch==2.8.0, pytorch-triton-rocm, torchvision, torchaudio, torchao==0.13.0, xformers)

    references/llms-full.md:1532 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (unsloth, unsloth-zoo)

    references/llms-full.md:1546 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (git+https://github.com/unslothai/unsloth)

    references/llms-full.md:1548 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (diffusers, as, well, if, you, get …)

    references/llms-full.md:3444 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (unsloth, vllm)

    references/llms-full.md:3546 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (vllm)

    references/llms-full.md:4563 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (uv)

    references/llms-full.md:5308 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (transformers==4.56.2, tokenizers)

    references/llms-full.md:5316 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (huggingface_hub, hf_transfer)

    references/llms-full.md:6095 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (torch==2.8.0, torchvision, torchaudio)

    references/llms-full.md:7375 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (https://github.com/sgl-project/whl/releases/download/v0.3.2/sgl_kernel-0.3.2+cu128-cp39-abi3-manylinux2014_x86_64.whl)

    references/llms-full.md:7386 · skill_risk:S2

  • high

    Tells Claude to install unpinned packages (https://github.com/sgl-project/whl/releases/download/v0.3.2/sgl_kernel-0.3.2-cp39-abi3-manylinux2014_x86_64.whl)

    references/llms-full.md:7390 · skill_risk:S2

medium · 3

  • medium

    Tells Claude to install packages (unsloth==2025.1.5)

    references/llms-full.md:946 · skill_risk:S3

  • medium

    Tells Claude to install packages (trl==0.22.2)

    references/llms-full.md:5317 · skill_risk:S3

  • medium

    Tells Claude to install packages (sgl-kernel==0.3.2)

    references/llms-full.md:7382 · skill_risk:S3

Versions

Version history

Each distinct content we have seen, newest first, with the verdict it got at the time.

One version so far: v1 (1.0.0), seen 2026-11-03.

See and control which skills your people use

Inside your organization, Waxell scans each skill and plugin before anyone installs it, approves what passes and publishes it to a private marketplace, and puts every MCP tool call behind one governed URL.

Book a demo

30 minutes. No commitment.

SOC 2 Type II attested · ISO/IEC 27001:2022 certified · HIPAA with a BAA · aligned to the EU AI Act and NIST AI RMF

More research skills and plugins

See all