social-content

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTIONNO_CODE
Full Analysis
  • PROMPT_INJECTION (LOW): The skill provides a framework for 'Reverse Engineering Viral Content' in 'references/reverse-engineering.md' which involves analyzing untrusted external data. This creates a surface for indirect prompt injection.\n
  • Ingestion points: The skill suggests collecting and analyzing 500-1000+ posts from external creators (references/reverse-engineering.md, Step 2 and 3).\n
  • Boundary markers: Absent; there are no instructions provided to separate the scraped data from the agent's core logic or to ignore embedded commands.\n
  • Capability inventory: Reasoning and text generation; the agent uses the data to extract patterns and 'Codify playbook' (references/reverse-engineering.md, Step 4).\n
  • Sanitization: Absent; no sanitization or validation of the input text is mentioned.\n- NO_CODE (SAFE): The skill consists entirely of Markdown files and does not contain any executable scripts, Python packages, or automated command execution.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 06:44 AM