revid-research-make-and-post
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on untrusted data retrieved from external social media platforms, creating a potential surface for indirect prompt injection.
- Ingestion points: The agent retrieves transcripts and hooks from viral videos using the
search_viral_videostool. - Boundary markers: The instructions lack specific delimiters or warnings to the agent to ignore any commands or malicious instructions that might be embedded within the retrieved video transcripts.
- Capability inventory: The skill possesses significant capabilities, including video rendering (
render_video), listing connected social media accounts (list_social_accounts), and publishing or scheduling content to public social media profiles (publish_now,schedule_publish). - Sanitization: There is no evidence of sanitization or filtering of the external transcripts before they are used as references for drafting new video scripts.
- [EXTERNAL_DOWNLOADS]: The skill depends on an external Revid MCP server located at
https://www.revid.ai/api/mcpto perform its core video generation and social media management functions.
Audit Metadata