Security audit

professional-pdf-translator

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent PDF translation workflow that reads user-selected documents and writes expected translation/glossary outputs, with no hidden or unrelated behavior found.

Install only if you are comfortable letting the agent read the selected PDF and glossary files and write the chosen DOCX, temporary text, and dynamic glossary files. For confidential documents, choose a controlled output directory and avoid pointing the output or glossary path at existing important files.

SkillSpector

Vulnerability Patterns

MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: The skill clearly performs filesystem reads and writes via local PDF input, DOCX output, temporary text files, and a continually appended dynamic glossary, yet it declares no permissions. This creates a transparency and consent problem: a user or platform may not understand that the skill can modify local files, increasing the risk of unintended data changes or misuse if invoked with sensitive paths.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The workflow creates and continually appends to a dynamic glossary file near the source PDF, but this side effect is not prominently disclosed as a write operation with persistence over time. That omission can lead to silent modification of user directories, leakage of extracted sensitive terms into a new artifact, and surprises when processing confidential documents.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill incrementally writes to the specified DOCX during processing, but the description does not clearly warn that the file will be modified page by page rather than produced only at the end. This can cause accidental overwrite/corruption of an existing document, partial-output exposure if the run fails midway, and unsafe assumptions about atomicity or recoverability.

Missing User Warnings

Low

Confidence: 89% confidence
Finding: The script accepts an arbitrary --file path and will create parent directories, create the file if missing, and append attacker-controlled content to that path with no restriction or confirmation. In an agent workflow, this can be abused to overwrite or pollute sensitive user files or application data if an untrusted prompt, tool input, or compromised workflow controls the output path.

Static analysis

No suspicious patterns detected.