x-post-research
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill performs network requests to the FxTwitter API to collect public posts. Instructions specifically mandate that no credentials, tokens, or cookies be sent, ensuring only public data is accessed.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text from external social media posts, creating a potential surface for indirect injection.
- Ingestion points: Public X posts and their associated metadata retrieved from the web.
- Boundary markers: The skill requires the agent to categorize all findings into specific evidence levels (A-D) and explicitly attribute claims to their sources, preventing external text from being processed as instructions.
- Capability inventory: File system access for writing research notes and raw JSON data, and network access for API queries and URL resolution.
- Sanitization: The workflow requires cross-referencing all social media claims with official documentation and code repositories before final reporting.
Audit Metadata