social-content
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- PROMPT_INJECTION (LOW): The skill's 'Reverse Engineering' framework in
references/reverse-engineering.mdcreates a surface for indirect prompt injection. * Ingestion points: Scraped content from platforms like LinkedIn and Twitter/X as part of the 500-1000 post collection process. * Boundary markers: Absent; the instructions do not suggest using delimiters or system instructions to treat the scraped content as data rather than instructions. * Capability inventory: The agent generates social media posts, threads, and strategic playbooks based on the ingested content. * Sanitization: Absent; no mention of filtering or sanitizing scraped text before analysis. - NO_CODE (SAFE): The skill contains only Markdown files and no executable scripts or binaries, which significantly reduces the risk of malicious local execution or unauthorized system access.
Audit Metadata