WriteStory

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The skill mandates executing an immediate curl POST to a local /notify endpoint before doing anything else, which is a side-effectful network action unrelated to the stated fiction-writing purpose and therefore a hidden/out-of-scope instruction.

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow for WriteStory’s Interview/BuildBible/WriteChapter/Revise ingests user-provided free text written into the user’s own project files (e.g., interview.md, saved notes/seed material, and existing chapters) and has the LLM read it to derive and rewrite story elements, so outsider-authored text can be submitted via the user’s inputs/files.