16-ai-llm-security
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEPROMPT_INJECTIONOBFUSCATIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill includes various prompt injection examples and test payloads in SKILL.md and scripts/prompt_injection_tester.py, such as 'Ignore all previous instructions' and 'SYSTEM OVERRIDE' patterns. These are explicitly defined as part of a testing corpus for auditing third-party systems and do not affect the agent's own instructional integrity.
- [OBFUSCATION]: The scripts contain examples of obfuscated payloads for testing purposes, including Base64-encoded strings ('SWdub3JlIHJ1bGVz...') and zero-width character insertion ('Ignore'). These are identified as test vectors for security assessments and are not used to conceal the skill's own operations.
- [EXTERNAL_DOWNLOADS]: The skill identifies dependencies on standard, well-known libraries such as requests, pyyaml, and rich. It also references established security tools and frameworks from trusted sources like NVIDIA, MITRE, and OWASP.
Audit Metadata