design-md-generator
Pass
Audited by Gen Agent Trust Hub on Apr 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
scripts/extract_css.pyvia shell command to process website data provided by the user. - [EXTERNAL_DOWNLOADS]: The skill requires the
requestslibrary to fetch HTML and CSS files from external websites. - [DATA_EXFILTRATION]: The Python script performs network operations to non-whitelisted, user-specified domains to retrieve site assets for analysis.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted data from external websites.
- Ingestion points: Website content (HTML/CSS) is fetched via
extract_css.pyfrom URLs provided in user prompts. - Boundary markers: No explicit delimiters or instructions are used to prevent the agent from obeying instructions potentially embedded in the scraped site data.
- Capability inventory: The agent has the ability to execute shell commands (running the script) and perform network requests (via the script).
- Sanitization: The
extract_css.pyscript utilizes regular expressions to extract specific CSS properties, which provides a layer of data structuralization that helps filter out non-design content.
Audit Metadata