free-translation-api
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill makes network requests to well-known public LibreTranslate instances, including libretranslate.com, translate.argosopentech.com, and translate.terraprint.co, to perform translation, language detection, and metadata retrieval.
- [COMMAND_EXECUTION]: The skill uses shell commands like curl and jq to interact with APIs and system tools. It includes scripts that read local files (e.g., document.txt, README.md) and write translated output to the filesystem.
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface where the content of local files is directly interpolated into shell commands and API request bodies without sanitization.
- Ingestion points: Local files such as document.txt, content_en.txt, and README.md are read using cat and read in the translate_text, batch_translate, and translate_markdown_document examples.
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the provided scripts.
- Capability inventory: The skill utilizes curl for network communication, cat/read for file reading, and shell redirection (>) for file writing.
- Sanitization: File contents are used directly in string interpolation for shell commands and JSON bodies without escaping, which could lead to command malformation or indirect manipulation of the agent's logic if the translated content is subsequently followed as instructions.
Audit Metadata