storm-research

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or data exfiltration techniques were identified within the skill's instructions or logic.
  • [COMMAND_EXECUTION]: Shell command execution is properly limited through the allowed-tools frontmatter configuration, which restricts the Bash tool to a safe subset of utilities (mktemp, mkdir, date). This prevents arbitrary command injection.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by fetching untrusted external content from the web to generate personas and drive research simulations. This is a design requirement for the skill's primary research purpose.
  • Ingestion points: External content retrieved via WebSearch and WebFetch (SKILL.md).
  • Boundary markers: The skill does not explicitly define delimiters to separate retrieved web data from system-level instructions.
  • Capability inventory: The agent has access to file-writing tools, web search, and the ability to load other skills (SKILL.md).
  • Sanitization: Includes basic citation hygiene (e.g., stripping inline tags) but does not define validation or filtering for malicious instructions within web snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 03:06 AM
Security Audit — agent-trust-hub — storm-research