asu-resume
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data such as user resumes, PDFs, and project documentation to generate HTML content. This presents a potential attack surface where instructions embedded in a resume could influence the agent's behavior. However, the skill implements a 'claim-evidence-ledger' to track verified information and uses explicit markers like
【待补：指标/职责/链接】for unconfirmed data, which serves as a mitigation strategy. - [EXTERNAL_DOWNLOADS]: The skill downloads brand icons and SVG assets from LobeHub's official repositories (
@lobehub/iconsand@lobehub/icons-static-svg). These are well-known resources used for UI enhancement and are fetched to represent company logos in the generated resumes. - [DYNAMIC_EXECUTION]: The generated resume files include a functional toolbar for client-side editing, formatting, and local storage persistence. This involves the generation of HTML with embedded JavaScript to provide features like undo/redo and A4 page balancing, scoped to the user's local browser environment.
Audit Metadata