social-media-analyzer

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Prompt Injection (SAFE): No instructions found that attempt to override system prompts, bypass safety filters, or extract sensitive model information. The instructions are focused exclusively on the intended domain of social media analysis.
  • Data Exposure & Exfiltration (SAFE): The scripts do not contain hardcoded credentials, API keys, or access to sensitive local paths. No network operations (requests, urllib) are present in the code.
  • Obfuscation (SAFE): All markdown and Python content is provided in clear text with no evidence of Base64 encoding, zero-width characters, homoglyphs, or other obfuscation techniques.
  • Unverifiable Dependencies & Remote Code Execution (SAFE): The skill does not download external packages or execute remote scripts. No dynamic execution functions like eval() or exec() are utilized.
  • Indirect Prompt Injection (SAFE): While the skill processes external campaign data (JSON/CSV), it lacks the necessary capabilities (such as file-writing or network access) to be exploited via data poisoning. The logic is restricted to arithmetic and string formatting.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:32 PM