li-progressive-case-story
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes a local hidden directory
~/.li/to store a file named个人故事库.md. This file serves as a persistent database for user stories, allowing the agent to retrieve previously shared context. This behavior is consistent with the stated purpose of creating a personal story library and aligns with the author's namespace. - [PROMPT_INJECTION]: The skill ingests data from the local
~/.li/个人故事库.mdfile to generate scripts. This constitutes a surface for indirect prompt injection. Ingestion point:~/.li/个人故事库.md. Boundary markers: Absent. Capability inventory: No shell execution or network operations detected across scripts. Sanitization: None. While the surface exists, the risk is negligible as the data source is user-controlled local storage and the skill lacks exploitable capabilities. - [SAFE]: The skill's instructions are focused on structured interviewing and text transformation. There are no patterns suggesting data exfiltration, obfuscation, or unauthorized privilege escalation.
Audit Metadata