storm-research

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it fetches and processes external data from the web.
  • Ingestion points: Phase 1 (expert lenses) and Phase 4b (citation verification) utilize web search and fetch tools to ingest external content (SKILL.md).
  • Boundary markers: The skill does not implement delimiters or specific 'ignore' instructions for the fetched content when providing it to the sub-agents.
  • Capability inventory: The skill has access to the Agent tool for spawning sub-agents, the Write tool for file creation, and the ability to execute system commands (open, xdg-open, start).
  • Sanitization: No explicit sanitization or validation of the untrusted web content is performed before processing (SKILL.md).
  • [COMMAND_EXECUTION]: The skill uses shell commands to automatically open the research briefing. It uses a filename derived from user input in the research topic, which could potentially be manipulated if the 'kebab-case' transformation is not strictly enforced (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 04:41 AM
Security Audit — agent-trust-hub — storm-research