html-to-markdown-fetching-and-converting-urls
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates fetching and processing data from arbitrary external URLs using the
--urlflag. This creates a surface where instructions embedded in the source HTML could potentially influence the AI agent's behavior. \n - Ingestion points: The
html-to-markdown --url <URL>command inSKILL.mdallows the agent to ingest external content.\n - Boundary markers: The instructions do not currently specify the use of delimiters or warnings for the agent to ignore embedded instructions in the fetched content.\n
- Capability inventory: The tool can write output to the file system using the
-oor--outputflag (e.g.,page.md).\n - Sanitization: The tool provides robust sanitization through the
--preprocessflag and aggressiveness presets (minimal,standard,aggressive), which strip scripts, forms, and navigation elements.\n- [EXTERNAL_DOWNLOADS]: The documentation includes installation commands for various environments, including Brew, NPM, PyPI, Cargo, NuGet, and Maven. All packages (such as@xberg-io/html-to-markdown,html-to-markdown, andXbergIo.HtmlToMarkdown) and source repositories are officially maintained by the skill author (xberg-io) and are consistent with the tool's intended purpose.\n- [COMMAND_EXECUTION]: The skill instructs the agent on how to execute thehtml-to-markdownCLI tool to perform conversions, generate shell completions, and produce JSON output. These commands are necessary for the primary function of the skill and do not involve unauthorized privilege escalation or persistence mechanisms.
Audit Metadata