afa-social
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is composed of markdown files providing instructions and reference data for social media marketing. It does not include any executable scripts, binary files, or system commands, eliminating the possibility of local or remote code execution.
- [SAFE]: No sensitive data exposure or exfiltration patterns were identified. The skill references local data files like 'products.md' and 'learnings.jsonl' within its defined operational scope and does not perform network operations to untrusted domains.
- [SAFE]: No obfuscation techniques or hidden commands were found. Text analysis for Base64, hex-encoding, and hidden Unicode characters returned no suspicious results.
- [SAFE]: The skill includes instructions to prevent the use of unauthorized user content and provides clear operational boundaries, mitigating risks associated with processing external data (Indirect Prompt Injection). Ingestion points are limited to brand data and UGC; capabilities are restricted to text generation and logging; and boundary markers are explicitly defined in the startup protocol.
Audit Metadata