Security Audit

huggingface-datasets

github.com/huggingface/skills

AI SkillCommit 80f9fa530e46

CRITICAL

Scanned 14 days ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

huggingface-datasets received a trust score of 31/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.

SkillShield's automated analysis identified 3 findings: 1 critical, 2 high, 0 medium, and 0 low severity. Key findings include File read + network send exfiltration, Sensitive path access: AI agent config, Data Exfiltration of Sensitive Agent Session Traces.

The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.

Last analyzed on October 1, 2026 (commit 80f9fa53). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.

Layer Breakdown

Manifest Analysis

70%

Static Code Analysis

85%

Dependency Graph

100%

LLM Behavioral Safety

85%

Behavioral Risk Signals

Network Access

2 findings

Filesystem Write

3 findings

Shell Execution

2 findings

Excessive Permissions

2 findings

Security Findings3

Severity	Finding	Layer	Location
CRITICAL	File read + network send exfiltration AI agent config/credential file access Remove access to sensitive files not required by the skill's stated purpose. SSH keys, cloud credentials, and browser data should never be read by skills unless explicitly part of their declared functionality.	Manifest	skills/huggingface-datasets/SKILL.md:94
HIGH	Sensitive path access: AI agent config Access to AI agent config path detected: '~/.claude/'. This may indicate credential theft. Verify that access to this sensitive path is justified and declared.	Static	skills/huggingface-datasets/SKILL.md:94
HIGH	Data Exfiltration of Sensitive Agent Session Traces The skill instructs the agent to locate and upload local agent session traces (from Claude Code, Codex, or Pi Agent) to Hugging Face. These session traces contain highly sensitive information, including command history, file paths, tool outputs, private code, and potentially active secrets or API keys. Uploading these files to an external platform poses a severe data exfiltration and privacy risk. Remove the instructions and commands that suggest uploading local agent session traces or directories containing sensitive session history (such as `~/.claude/projects`) to Hugging Face.	LLM	SKILL.md:69

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/e905c835456d0e42.svg)](https://skillshield.io/report/e905c835456d0e42)