lov-search-twitter
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text content from multiple external sources, including third-party Twitter renderers (fxtwitter.com, vxtwitter.com) and public archives (Internet Archive, Common Crawl). This content is incorporated into the agent's context and reports, which could potentially contain malicious instructions intended to influence the agent's behavior in subsequent turns. \n
- Ingestion points: scripts/search_twitter.py fetches HTML and JSON payloads from external mirror sites and archival services. \n
- Boundary markers: The skill does not implement specific delimiters or 'ignore embedded instructions' markers when presenting recovered post text to the agent or user. \n
- Capability inventory: The skill possesses network access capabilities (scripts/search_twitter.py) and local file system write capabilities (scripts/profile_store.py). \n
- Sanitization: Content processing is limited to basic HTML unescaping and JSON parsing without instruction-level sanitization.
Audit Metadata