Security audit

python-packaging

Security checks across malware telemetry and agentic risk

Overview

This is a straightforward Python packaging guide, but users should recognize that its publish examples can upload packages to public registries.

Install only if you want packaging and release guidance. Before running uv publish or enabling the GitHub Actions examples, confirm the target registry, credentials, package contents, version, and whether the release will become public. Consider testing on TestPyPI first.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Intent-Code Divergence

Low

Confidence: 84% confidence
Finding: The documentation labels project structure snippets as examples of directory layouts, but the embedded and adjacent verification lines instruct users to run `pytest -v` or a command with `--help`, which does not meaningfully verify those structure examples. This is an intent-documentation mismatch within the file rather than a code-execution issue, because the stated verification action contradicts what the content actually presents.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The manifest trigger list contains very broad words such as "python", "packaging", and "pip", plus a general phrase about publishing or build configuration. These terms overlap with common development discussions and the file does not provide clear scope limits or negative examples, making accidental activation more likely.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The markdown quick-start includes `uv publish`, which can upload a package to an external registry, but the skill text does not warn users that this may publicly distribute code or metadata. For a markdown skill description, this is a user-impacting action affecting data and system integrity that should be disclosed clearly.

Missing User Warnings

Low

Confidence: 78% confidence
Finding: This markdown file describes an automated publishing workflow that uses a PyPI token and runs `uv publish` on release events. While it notes that tokens should be stored as GitHub secrets, it does not explicitly warn users that publishing is automated and can push artifacts to a public package registry when a release is published.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: This markdown file includes `uv publish` examples for TestPyPI and PyPI, which can transmit package contents and metadata to external registries and may publicly release artifacts. The section provides no warning or cautionary note about the external publication effect, making the behavior easy to invoke without explicit user awareness.

Overly Broad Trigger

Low

Category: Trigger Abuse
Confidence: 70% confidence
Finding: Overly Broad Trigger: 'uv' is too short and may match unintended inputs

VirusTotal

57/57 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.