short-drama-produce

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow reads from and edits user-authored Markdown files (图片提示词.md, 视频提示词.md, 分镜.md, 剧本.md) which can contain free text written by project users/creators.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (medium risk: 0.30). The skill executes media production tasks (images, videos, TTS, and music) via external adapters (such as Seedance, MiniMax, and GPT Image 2) which directly submit jobs and incur charges at submission time (SKILL.md:147-150). This represents a direct financial operation (purchasing/generating paid API service calls). Access control for these operations relies on a mandatory hard gate workflow requiring explicit creator preview (prepare), explicit creator confirmation (confirm), and an execution tool (run) that validates the confirmation fingerprint before calling the adapter (SKILL.md:43-53, 86).