creating-characters
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill workflow uses user-controlled input to determine file names for output storage.
- Ingestion points: User-provided character titles (
{タイトル}) inSKILL.md(Step 9). - Boundary markers: Absent; the workflow does not instruct the agent to sanitize or validate the title string before use.
- Capability inventory: The skill performs file-write operations to the local filesystem (
output/character/). - Sanitization: Absent; there are no instructions to prevent path traversal characters (e.g.,
../) in the user-provided title. - Relative Path Reference (LOW): The workflow references a file outside its immediate directory (
../ANTI_AI_STYLE.md). While likely intended for shared resources, accessing parent directories can be a sign of poor isolation in some agent environments.
Audit Metadata