web-content-downloader
Pass
Audited by Gen Agent Trust Hub on Mar 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses terminal commands including
curlfor network requests,mkdirfor directory creation, andmvfor file renaming operations during the image processing workflow. - [EXTERNAL_DOWNLOADS]: Fetches content from user-specified URLs and downloads images to the local environment. It utilizes Jina Reader (r.jina.ai), a well-known service for web-to-markdown conversion, for its primary fetching method.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes arbitrary content from the internet. Ingestion points: External web content fetched via
curlor DuckDuckGo Search (SKILL.md). Boundary markers: Absent; there are no instructions to the agent to treat downloaded content as untrusted or to use delimiters. Capability inventory: Shell command execution (curl), directory management (mkdir), and file system write access (saving Markdown files and images). Sanitization: Absent; the content is retrieved and saved without validation or escaping of potential malicious instructions.
Audit Metadata