industrial-ai-research
Pass
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill ingests untrusted data from external research papers via WebSearch and WebFetch (Ingestion points: SKILL.md, Phase 3). It currently lacks explicit boundary markers to isolate this fetched content. Capability inventory includes Read, Glob, Grep, WebSearch, and WebFetch (SKILL.md), providing limited exploitable tools. Sanitization is implemented through a mandatory Verification and Triage phase (SKILL.md, Phase 4) that removes low-quality sources.
- [NO_CODE]: The skill contains no executable scripts or binary files, consisting entirely of Markdown and YAML instructions. This architecture limits the risk of direct malicious code execution or sophisticated obfuscation.
Audit Metadata