Survival Analysis (KM)
Security checks across malware telemetry and agentic risk
Overview
This appears to be a local survival-analysis tool with no evidence of hidden or malicious behavior, but users should handle clinical data and Python dependencies carefully.
This skill looks coherent for local Kaplan-Meier survival analysis. Before installing or running it, use a dedicated Python environment, consider pinning dependency versions, and make sure any clinical input data and generated reports are stored only in approved locations.
VirusTotal
65/65 vendors flagged this skill as clean.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing later package versions could change behavior or inherit upstream package risk.
The skill depends on third-party Python packages without version pins. This is common and purpose-aligned for a scientific analysis script, but it leaves exact package versions to the install environment.
lifelines matplotlib numpy pandas seaborn
Install in a dedicated Python environment and consider pinning known-good package versions for reproducible clinical analyses.
The tool can read the selected input file and create or overwrite expected result files in the selected output location.
The script reads a user-specified CSV and writes results to a user-specified output directory. This is necessary for the tool's purpose, but it means the user or agent controls local file paths.
parser.add_argument('--input', '-i', required=True, help='Input CSV file path')
parser.add_argument('--output', '-o', required=True, help='Output directory for results')Use explicit, project-specific input and output paths, and avoid pointing the output directory at locations containing important files.
Generated plots, CSV summaries, and reports may reveal information derived from clinical datasets.
The documented use case includes clinical survival datasets, which may contain sensitive patient or study data. The outputs are local analysis artifacts rather than agent memory or network sharing, but users should still treat them as sensitive.
clinical and biological survival data analysis
Use de-identified datasets where possible and store generated outputs only in approved project locations.