persona-model-trainer

Fine-tune any HuggingFace instruction-tuned model (Gemma 4, Qwen 3, Llama, Phi, Mistral, and more) on persona data from anyone-skill. Produces a self-contain...

This skill largely does what it says (local persona fine-tuning) but you should not install blindly. Before running it: 1) Review training data for PII and remove/redact sensitive content (prepare_data.py helps but is not perfect). 2) Inspect scripts/train.py, scripts/export.py and any generated root-level wrappers for unexpected network calls or hardcoded remote endpoints. 3) Note the skill assumes external tools (Python 3.11+, torch/peft/bitsandbytes, ollama, llama.cpp convert script, vLLM, mlx-lm, Unsloth) — install them from official sources and audit any third-party wheels. 4) Be cautious with the optional autoresearch integration: it edits project scripts and runs training loops autonomously — run it only in an isolated environment and after reading .agents/skills/autoresearch/SKILL.md. 5) If you plan to publish or push adapter weights, obtain explicit consent from any people whose data was used and ensure you configure HuggingFace (or other) auth tokens securely. 6) If you see any 'ignore previous instructions' or system-override text in the SKILL.md or generated notebooks, treat it as suspicious and remove or sanitize before running. If you want, I can (a) list the exact locations of prompt-injection-like strings inside SKILL.md and scripts, or (b) produce a checklist of binaries and environment setup commands to run in a safe sandbox.

Type: OpenClaw Skill Name: persona-model-trainer Version: 0.3.3 The persona-model-trainer skill bundle is a legitimate and well-documented toolset for fine-tuning HuggingFace models on persona data. It provides a complete pipeline including data preparation (prepare_data.py), environment validation (check_env.py), multi-backend training (train.py), and model export (export.py). The bundle demonstrates strong security and privacy awareness by including explicit instructions in SKILL.md to treat training data as untrusted to prevent prompt injection, implementing a PII scanner in prepare_data.py, and providing comprehensive documentation on data handling (privacy.md). All high-risk capabilities, such as shell execution and network access for model uploads, are strictly aligned with the stated purpose of model training and deployment.