account-backup
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill is designed to access and aggregate sensitive personal information (PII) from x.com, specifically including account bookmarks, following lists, and profile data, which are then exported to a JSON format.
- [COMMAND_EXECUTION]: The instructions direct users to paste and execute JavaScript code directly in the browser console (DevTools) while authenticated on x.com. This pattern, known as Self-XSS, is a high-risk activity that can be used to bypass security boundaries and potentially steal session cookies or authentication tokens.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external, untrusted content from social media profiles which serves as a vulnerability surface. 1. Ingestion points: Scraped DOM content from tweets, profiles, and follower lists (SKILL.md). 2. Boundary markers: None provided to delimit the scraped content or instruct the agent to ignore embedded instructions. 3. Capability inventory: Browser console execution, automated JSON downloads, and API route navigation. 4. Sanitization: No sanitization or validation of the scraped data is mentioned.
Audit Metadata