niche-research
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of a single markdown file (SKILL.md) that provides instructional guidelines for niche research. No executable scripts, shell commands, or network operations are present. The instructions focus on structured analysis and language acquisition for business development. No prompt injection, obfuscation, or data exfiltration patterns were detected.\n- [INDIRECT_PROMPT_INJECTION]: The research framework involves gathering data from public platforms like Reddit and industry forums. This ingestion of untrusted data is a potential attack surface; however, the skill does not grant the agent high-privilege capabilities (such as file-system writes or unauthorized network access) that would make this surface exploitable. The primary output is a natural language report.\n
- Ingestion points: Public forums, subreddits, and industry websites (SKILL.md).\n
- Boundary markers: None explicitly defined in the instructions for the agent during the search process.\n
- Capability inventory: None; the skill does not request access to high-privilege tools or scripts.\n
- Sanitization: Not applicable as the skill does not involve data processing logic beyond reporting findings.
Audit Metadata