write-tweet
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from user inputs and external URLs, creating a surface for potential indirect prompt injection attacks.
- Ingestion points: The skill instructions in
SKILL.mdallow the agent to fetch content from user-provided links ("用户给的是链接时先取全文") and process user-supplied text fragments. - Boundary markers: The instructions do not specify the use of delimiters or specific "ignore instructions" warnings when handling the external "material" (材料).
- Capability inventory: The skill generates public-facing text (tweets) and interacts with a sub-agent tool (
/writing-ai-check). - Sanitization: No explicit sanitization or filtering is defined for the content fetched from external links or provided by the user before it is interpolated into the workflow.
Audit Metadata