web-article-translator
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (HIGH): The skill processes untrusted data from external URLs and possesses file-write capabilities, creating a significant attack surface.
- Ingestion points: Content fetched from external URLs via the
webReadertool. - Boundary markers: Absent. There are no instructions to the agent to treat the fetched Markdown content strictly as data or to ignore embedded instructions.
- Capability inventory: File system write operations (
translated-{title}.md). - Sanitization: Absent. The skill instructions do not require the agent to filter or escape the fetched content before processing or saving.
- Command Execution (LOW): While the skill uses a tool called
webReader, there is no evidence of arbitrary shell command execution or untrusted script execution. The risk is primarily logical rather than technical code execution.
Recommendations
- AI detected serious security threats
Audit Metadata