social-content
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is composed entirely of Markdown documentation and text-based templates. No Python code, shell scripts, or external dependencies are included, which eliminates the risk of direct remote code execution or unauthorized command execution.\n- [PROMPT_INJECTION]: The 'Reverse Engineering' framework outlined in
references/reverse-engineering.mdinvolves the ingestion and analysis of 500-1000+ external social media posts. This methodology introduces a surface for indirect prompt injection, as the agent is instructed to process a high volume of untrusted content that could contain hidden instructions.\n - Ingestion points: External social media posts gathered through manual collection or third-party tools as suggested in
references/reverse-engineering.md.\n - Boundary markers: No specific delimiters or instructions for the agent to ignore potentially malicious embedded content within the ingested data are present.\n
- Capability inventory: The skill does not define any scripts, subprocesses, network requests, or file-writing operations across its files.\n
- Sanitization: No sanitization or data validation steps are mentioned in the content analysis framework.
Audit Metadata