social-content
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is composed entirely of informational documentation and templates. No security threats, obfuscation, or hidden malicious patterns were found across any of the files.
- [NO_CODE]: No executable logic, such as Python, JavaScript, or shell scripts, is provided with this skill. It relies purely on natural language instructions for the AI agent.
- [DATA_EXFILTRATION]: There are no network operations, API calls, or data exfiltration vectors present. The skill references a local file
.claude/product-marketing-context.mdfor context gathering, which is a standard and secure practice for agent personalization within the local environment. - [PROMPT_INJECTION]: The skill outlines a framework for analyzing scraped social media content (Step 2 in
references/reverse-engineering.md), which constitutes an ingestion point for untrusted data. However, the skill lacks any 'high-risk' capabilities, such as shell access or dynamic code execution, and contains no inventory of tools that could be exploited via indirect prompt injection. Boundary markers and sanitization are absent but unnecessary given the skill's lack of executable actions.
Audit Metadata