spark
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-supplied text ('inspirations') stored in a local file (
灵感本.md) to generate outlines and drafts. This design creates an attack surface for indirect prompt injection, where malicious instructions hidden within saved notes could be interpreted as agent commands during the generation phase. - Ingestion points: Processes data from the local file
灵感本.mdand direct user input strings. - Boundary markers: The skill instructions do not define delimiters or specific warnings to distinguish user content from agent instructions during interpolation.
- Capability inventory: The skill is capable of writing to the local file system and performing multi-step content generation.
- Sanitization: No validation, escaping, or filtering of user-provided content is implemented before it is used to drive the drafting workflow.
Audit Metadata