educational-content-and-how-to

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified. The skill consists entirely of instructional Markdown files and evaluation data intended for content planning.
  • [NO_CODE]: The skill does not include any executable scripts, binaries, or configuration files that could be used for code execution or persistence.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to its design of processing external research data to generate social media posts.
  • Ingestion points: Reads and processes data from audience-research, brand-profile, and content-pillars (SKILL.md).
  • Boundary markers: Absent; there are no specific delimiters or instructions provided to the agent to distinguish between its core instructions and potentially adversarial content within the research data.
  • Capability inventory: The skill routes generated copy to format writers which connect to publishing tools like WoopSocial (references/scope-and-connections.md).
  • Sanitization: Absent; the skill does not specify any validation, filtering, or escaping of the research data before it is incorporated into the lesson structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 02:27 PM
Security Audit — agent-trust-hub — educational-content-and-how-to