event-booth-experience

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data through a structured 11-question interview to generate event floor briefs. This data ingestion, combined with the use of persistent memory for edition records, creates an indirect prompt injection surface.\n
  • Ingestion points: User answers to interview questions in SKILL.md and edition-specific data stored in persistent memory.\n
  • Boundary markers: None defined; there are no instructions to the agent to treat external data as untrusted or to ignore embedded commands.\n
  • Capability inventory: The skill does not possess tools for network access, file system modification, or command execution.\n
  • Sanitization: No validation or sanitization of user input is performed.\n- [NO_CODE]: The skill consists solely of Markdown instructions and documentation. It does not include scripts, binaries, or references to third-party code packages.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 09:13 PM
Security Audit — agent-trust-hub — event-booth-experience