Ml Visualizer

Visual analysis and diagnostic tools to help machine learning model selection. ml-visualizer, python, anaconda, estimator, machine-learning, matplotlib.

This skill appears to be a simple local journal/logger for ML workflows and is internally consistent with that purpose. Before installing, consider: - The script writes logs to ~/.local/share/ml-visualizer by default. If you plan to record sensitive dataset paths or metadata, those will be stored as plain text on disk and included in exports. - SKILL.md claims an ML_VISUALIZER_DIR env var override, but the included script does not use it—so the data directory is hardcoded unless you modify the script. - There are minor metadata mismatches (SKILL.md name/version differs from registry metadata); these are likely sloppy packaging rather than malicious behavior. If you’re comfortable with local plain-text logs and/or you review/modify the script to change the data directory or tighten file permissions, this skill is coherent and low-risk. If you need the env-override behavior or stricter storage controls, ask the author for a corrected release or inspect and edit the script before use.

Type: OpenClaw Skill Name: ml-visualizer Version: 1.0.0 The skill is classified as suspicious due to a significant discrepancy between its stated purpose (ML visualization using matplotlib and anaconda) and its actual implementation, which is a simple bash script that only logs text strings to local files. Furthermore, the `search` command in `scripts/script.sh` is vulnerable to argument injection because it passes user input directly to `grep` without using the `--` delimiter, and the `_export` function is susceptible to JSON injection as it does not sanitize logged values before formatting them into a JSON structure.