bookmarks-management
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-generated content from X (Twitter) bookmarks, which could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: X/Twitter bookmark text and metadata processed by scripts referenced in
SKILL.md. - Boundary markers: The skill instructions do not provide delimiters or warnings to the agent to ignore instructions embedded within the bookmark data.
- Capability inventory: The skill interacts with the browser DOM to read and delete bookmarks and exports data to JSON/CSV files.
- Sanitization: There is no description of content sanitization or validation for the ingested bookmark data.
- [NO_CODE]: Several referenced script files (
scripts/scrapeBookmarks.js,src/bookmarkOrganizer.js,src/clearAllBookmarks.js,src/bookmarkManager.js) are missing from the skill definition, preventing a security audit of their internal logic and potential network or filesystem operations.
Audit Metadata