storm-research

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs automated web research by fetching content from external sources across multiple phases to synthesize its briefing. This involves ingesting untrusted data from the web.
  • [COMMAND_EXECUTION]: The skill utilizes platform-specific shell commands (open, xdg-open, or start) to automatically display the generated HTML report to the user. The command path is constructed using a kebab-case slug derived from the research topic.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted web data and incorporates it into agent context and generated artifacts.
  • Ingestion points: Web research and fetch activities in Phase 1 and citation verification in Phase 4 (SKILL.md).
  • Boundary markers: Absent. The sub-agent instructions do not include specific delimiters or 'ignore' warnings for the fetched text.
  • Capability inventory: The skill uses the Write tool to create files and shell execution to open them (SKILL.md).
  • Sanitization: Absent. The skill does not describe explicit sanitization or HTML escaping of fetched content before embedding it into the HTML report template.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 11:31 AM
Security Audit — agent-trust-hub — storm-research