x-create
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting content from user-provided template files without strict boundary delimiters. This could allow maliciously crafted templates to influence the agent's behavior during the content generation process.\n
- Ingestion points: The skill reads reference posts from the
assets/templates/directory during the content creation workflow (Step 1 inSKILL.md).\n - Boundary markers: The instructions lack explicit delimiters (e.g., XML-like tags) or directives to treat the template content solely as data.\n
- Capability inventory: The skill uses the agent's language model to generate tweets, threads, and replies, applying the persona and patterns identified in the templates (Steps 4 and 4.5 in
SKILL.md).\n - Sanitization: The skill does not perform validation or sanitization of the template files before incorporating them into the prompt.
Audit Metadata