deep-research
Pass
Audited by Gen Agent Trust Hub on Apr 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's HTML report template incorporates resources from well-known technology services, such as Tailwind CSS, ECharts (hosted on Cloudflare's CDN), Google Fonts, and Font Awesome. These dependencies are used for styling and data visualization in the final report.
- [DATA_EXFILTRATION]: The skill utilizes
web_searchandread_webpages_as_markdownto retrieve information from external websites. This activity is the core function of the skill and is intended for gathering research data. - [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted content from the web to generate reports.
- Ingestion points: Web content is ingested in Phase 3 of the
reference/research-workflow.mdusing theread_webpages_as_markdowntool. - Boundary markers: There are no explicit instructions or delimiters provided to isolate the untrusted web content from the agent's core instructions during the synthesis phase.
- Capability inventory: The agent can perform subsequent network searches and write to the local file system using
edit_fileduring the report generation phase. - Sanitization: The skill does not define specific sanitization or filtering logic for the content retrieved from the internet before it is incorporated into the final report output.
Audit Metadata