premise
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes user-supplied story ideas ("The Seed") and interpolates them into a structured output which is then written to a local file (
premise.md) usingfile:writepermissions. This creates a surface for indirect prompt injection, as the skill lacks instructions to delimit or sanitize the ingested data. - Ingestion points: User-provided initial ideas ("The Seed") described in the development process within
SKILL.md. - Boundary markers: The instructions do not specify any delimiters or safety warnings to ensure the agent ignores potentially malicious instructions embedded in the user's input.
- Capability inventory: The skill explicitly requests and uses
file:writepermissions to persist the generated premise to the file system (SKILL.md). - Sanitization: There are no procedures defined to escape or validate the external content before it is processed and stored.
Audit Metadata