OCR Space: Free OCR API

• Outcome: Review
• Latest audit: Sep 11, 2026 · 2:59 PM UTC
• Version: 1.0.0

Overview

This OCR skill appears to do what it claims, but it uploads images to OCR.space and has a local compressed-file overwrite/delete risk that users should review before installing.

Use this only for images you are comfortable sending to OCR.space. Avoid sensitive or regulated documents unless the external sharing is acceptable, and run it with ordinary user privileges in directories without valuable matching *_compressed.jpg files. The publisher should add explicit upload disclosure and replace the predictable compressed file path with a secure temporary file that is cleaned up reliably.

A.I.G

Vulnerability Patterns

• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Warning

• Location: ocr_space.py:67

Finding

Predictable Compression File Path Enables Arbitrary File Overwrite

Content

View full analysis

Remediation

View remediation

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Undeclared Tool Scope

Medium

• Category: MCP Least Privilege
• Confidence: 91% confidence

Finding

The skill advertises and documents a network-backed OCR capability but does not declare any explicit tool scope or permissions. This creates a transparency and governance gap: users or orchestration systems may invoke the skill without realizing it performs external network access, which can lead to unintended data exfiltration of image contents to a third-party service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 97% confidence

Finding

The skill description says it performs OCR but does not clearly warn that provided images are uploaded to the third-party OCR.space service. This is dangerous because users may pass screenshots, IDs, receipts, or other sensitive images under the assumption processing is local, resulting in unintended disclosure of personal, confidential, or regulated data to an external party.

Content

No source excerpt is available for this finding.

External Transmission

Medium

• Category: Data Exfiltration
• Confidence: 92% confidence

Finding

The code is explicitly designed to transmit image data to an external OCR API, which creates a real data egress path to a third party. In the context of an OCR skill, this is expected functionality, but it still becomes security-relevant when handling confidential images because sensitive content leaves the local trust boundary.

Content

Scanner excerpt · ocr_space.py (reported line 30)May include surrounding context.

python

import json
import base64

API_URL = "https://api.ocr.space/parse/image"
API_KEY = "helloworld"  # 免费测试 key

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 96% confidence

Finding

The script sends user-provided image contents to the third-party OCR.space service, but it does not present an explicit runtime warning or consent step about external data sharing. Images often contain sensitive personal or business information, so silent transmission to an external API can cause privacy and compliance issues in agent environments.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 80% confidence

Finding

The documentation and function signature set the default language to chs, making Simplified Chinese the default behavior unless the user supplies another language code. This is a locale/language preference baked in by default rather than chosen by the user.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low

• Category: Not specified by scanner
• Confidence: 83% confidence

Finding

The manifest description says the skill calls the OCR.space API to recognize text in images, which implies network OCR behavior. However, the implementation additionally rewrites oversized images into new local JPEG files as part of preprocessing, which is a file-modifying behavior not conveyed by the description.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.