unsloth-fine-tuning
Fast LLM fine-tuning with Unsloth - 2-5x faster training, 50-80% less VRAM. Use for single-GPU LoRA/QLoRA SFT, GRPO/RL reasoning training, vision/TTS fine-tuning, and GGUF export to Ollama/vLLM/llama.cpp. Supports 300+ models including Llama, Qwen, Gemma, DeepSeek, Mistral, Phi, and gpt-oss.
60
73%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Critical
Do not install without reviewing
Fix and improve this skill with Tessl
tessl review fix ./backend/cli/skills/ml-training/unsloth/SKILL.mdE005: Suspicious download URL detected in skill instructions.
What this means
Detected a suspicious URL in the skill instructions that could lead the agent to download and execute malicious scripts or binaries. This includes links to executables from untrusted sources, typosquatting of official packages, URL shorteners that obscure the destination, and personal file hosting services.
Why it was flagged
These URLs include typosquatted domains and direct install scripts/raw installers (e.g., .sh/.py) from non-standard or untrusted hosts, which are common vectors for distributing malware or running arbitrary code.
E006: Malicious code pattern detected in skill scripts.
What this means
Detected high-risk code patterns in the skill content — including its prompts, tool definitions, and resources — such as data exfiltration, backdoors, remote code execution, credential theft, system compromise, supply chain attacks, and obfuscation techniques.
Why it was flagged
The docs include high-risk remote code execution and supply‑chain patterns: a model tool-calling workflow exposes subprocess.run(..., shell=True) and exec(...) for arbitrary commands (and the runtime is invoked automatically via tool_choice="auto"), and there is a piped remote install script — all of which enable RCE/exfiltration if deployed or misused.
W007: Insecure credential handling detected in skill instructions.
What this means
The skill handles credentials insecurely by requiring the agent to include secret values verbatim in its generated output. This exposes credentials in the agent’s context and conversation history, creating a risk of data exfiltration.
Why it was flagged
The document contains examples that embed secrets directly into commands and code (e.g., docker -e JUPYTER_PASSWORD="mypassword" and model.push_to_hub_merged(..., token="...")), which encourages placing real passwords/API tokens verbatim into generated commands or code and therefore risks secret exfiltration.
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
The required runtime workflow shown in SKILL.md loads datasets from third-party Hugging Face repos via `load_dataset(...)` (outsider-authored free-form conversation text), which is then tokenized and becomes part of the LLM’s training context.
W012: Unverifiable external dependency detected (runtime URL that controls agent).
What this means
The skill fetches instructions or code from an external URL at runtime, and the fetched content directly controls the agent’s prompts or executes code. This dynamic dependency allows the external source to modify the agent’s behavior without any changes to the skill itself.
Why it was flagged
Installation instructions in the skill explicitly fetch-and-execute remote code (e.g. wget -qO- https://raw.githubusercontent.com/unslothai/unsloth/main/unsloth/_auto_install.py | python - and pip install from git+https://github.com/unslothai/unsloth.git / git+https://github.com/unslothai/unsloth_zoo.git), and there are git clone commands that pull and build remote code (https://github.com/ggml-org/llama.cpp), so these URLs constitute runtime external dependencies that will execute remote code if run.
- Repository
- synthetic-sciences/openscience
- Commit
1d182e9
- Audited
- Security analysis
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.