skills/haoxuanlithuai/awesome_cognitive_and_neuroscience_skills/cogsci-statistics/Gen Agent Trust Hub
cogsci-statistics
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The instructions focus entirely on domain-specific statistical guidance. No patterns suggesting behavioral overrides, safety filter bypasses, or instruction extraction were found.
- [DATA_EXFILTRATION]: The skill does not contain commands to access sensitive local files (e.g., .ssh, .env) or perform network operations to untrusted domains. References to GitHub are limited to the author's official repository for issue reporting.
- [REMOTE_CODE_EXECUTION]: All code snippets provided in the documentation use standard, well-established scientific and statistical libraries (such as lme4, statsmodels, mne, and nilearn). There are no commands for downloading or executing scripts from remote servers at runtime.
- [COMMAND_EXECUTION]: The provided code patterns are intended for user execution in data analysis environments (R/Python) and do not involve arbitrary shell command execution or system-level modifications.
- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or credentials were found. The skill does not instruct users to store secrets in unsafe locations.
Audit Metadata