content-atomizer
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion of external data. It instructs the agent to analyze and extract insights from sources like blog posts, videos, and reports. If these sources contain malicious instructions designed to hijack the agent's behavior, the agent may obey them as it lacks protective boundaries. \n
- Ingestion points: External source content (blogs, videos, podcasts, reports) specified in
SKILL.mdStep 1.\n - Boundary markers: Absent. The instructions do not provide delimiters or 'ignore' commands for embedded content.\n
- Capability inventory: The skill suggests using platform tools like
create_postandrate_contentin the 'Genfeed Integration' section.\n - Sanitization: Absent. No validation or filtering is applied to the input content.\n- [NO_CODE]: The skill contains no executable scripts, binaries, or logic-bearing files. It consists only of markdown instructions and metadata, which eliminates common attack vectors like malware or credential theft via code.
Audit Metadata