Mediapipe vision

• Outcome: Pass
• Latest audit: Aug 30, 2026 · 6:03 AM UTC
• Version: 1.0.0

Overview

The skill performs local MediaPipe computer-vision work as advertised, with privacy and dependency risks that users should understand before enabling webcam or saving outputs.

Install in an isolated Python environment, update or pin opencv-python to a patched version, and only allow webcam mode or persistent output files when you are comfortable capturing and storing the people, poses, faces, and surroundings in that media.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (8)

Missing User Warnings

Low

• Confidence: 84% confidence
• Finding: The README explicitly advertises live webcam processing but does not mention that activating this feature may access the user's camera and process real-time imagery. In a vision skill, this omission can weaken informed consent and increase privacy risk, especially if an agent invokes the skill on behalf of a user without clearly surfacing camera usage.

Missing User Warnings

Medium

• Confidence: 85% confidence
• Finding: The skill encourages live webcam processing but does not warn that it accesses the camera or that captured frames may include sensitive biometric and environmental information. In a vision skill, silent or insufficiently signposted camera use increases privacy risk because users may not realize they are enabling real-time collection of face, pose, and surrounding-scene data.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The skill documents exporting annotated media, JSON, CSV, and segmentation masks without warning that these files persist sensitive visual and biometric-derived data on disk. Saved landmarks, detections, and images can reveal identity, pose, gestures, or surroundings and may be later accessed, shared, or exfiltrated unintentionally.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The script can activate webcam capture when given the literal input 'webcam' without any explicit user-facing warning, confirmation, or prominent disclosure at the point of access. In an agent skill context, this increases privacy risk because a higher-level system could invoke the tool on a user's device and begin collecting live visual data unexpectedly.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# MediaPipe Vision Skill - Python dependencies
# Python 3.9-3.12 recommended (see mediapipe release notes for your platform).
mediapipe>=0.10.14
opencv-python>=4.8.0
numpy>=1.24.0

• Confidence: 91% confidence
• Finding: The dependency is specified with a lower bound only, so installation can resolve to different future versions over time. This creates a supply-chain and reproducibility risk because a later release could introduce breaking changes or a compromised package version without the project explicitly reviewing it.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# MediaPipe Vision Skill - Python dependencies
# Python 3.9-3.12 recommended (see mediapipe release notes for your platform).
mediapipe>=0.10.14
opencv-python>=4.8.0
numpy>=1.24.0

• Confidence: 95% confidence
• Finding: Using a non-pinned version for opencv-python is especially risky here because the allowed range includes 4.8.0, a version with known security advisories. In practice, environments may resolve to a vulnerable build, exposing users to downstream parsing risks from crafted image inputs.

Unpinned Dependencies

Low

• Category: Supply Chain

Content

# Python 3.9-3.12 recommended (see mediapipe release notes for your platform).
mediapipe>=0.10.14
opencv-python>=4.8.0
numpy>=1.24.0

• Confidence: 90% confidence
• Finding: An unpinned numpy dependency introduces reproducibility and supply-chain risk because future releases may change behavior or introduce vulnerabilities. While not immediately exploitable on its own, it weakens dependency control for the skill.

Known Vulnerable Dependency: opencv-python==4.8.0 — 2 advisory(ies): GHSA-qr4w-53vh-m672 (opencv-python bundled libwebp binaries in wheels that are vulnerable to CVE-2023); PYSEC-2023-183 (opencv-python versions before v4.8.1.78 bundled libwebp binaries in wheels that )

High

• Category: Supply Chain
• Confidence: 97% confidence
• Finding: The requirement opencv-python>=4.8.0 permits installation of opencv-python 4.8.0, which is documented as bundling vulnerable libwebp binaries in affected wheels. In a computer-vision skill that processes untrusted images and video, this context makes the issue more dangerous because malformed media could trigger vulnerable native code during decoding or preprocessing.

Static analysis

No suspicious patterns detected.