naiba-openai-hr

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill consists entirely of natural language prompt templates.
  • [NO_CODE]: The skill does not include any executable scripts, binary files, or external dependencies.
  • [PROMPT_INJECTION]: The skill includes prompts that ingest untrusted user data (e.g., survey responses, CSV datasets). This creates a surface for indirect prompt injection if the data being processed contains hidden commands.
  • Ingestion points: SKILL.md (Exit survey analysis, attrition analysis, and benchmarking prompts)
  • Boundary markers: Not used within the provided prompt templates
  • Capability inventory: None (no code or tool access is defined in this skill)
  • Sanitization: None provided
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 03:44 AM
Security Audit — agent-trust-hub — naiba-openai-hr