action-fight-prompt
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data such as character names, weapon types, and environmental descriptions to generate structured prompts for external video generation models.
- Ingestion points: The agent collects user input regarding duration, character details, weapon choices, and scene types in
SKILL.md(Step 1 of the workflow). - Boundary markers: The skill uses structured section headers (e.g.,
【资产锁定】,【场景】,【时序】) which provide implicit delimiters for the generated prompt segments. - Capability inventory: The skill is limited to text generation; it does not request or possess capabilities for network operations, file system writing, or code execution.
- Sanitization: There are no explicit instructions to sanitize or validate user-provided strings before they are interpolated into the final prompt template.
Audit Metadata