cross-post-video
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill follows secure design principles by including human-in-the-loop confirmation checkpoints (Steps 5 and 7) before the agent executes social media posts. It also performs validation of external media against platform rules before processing (Step 2).
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface through external media URLs, but it is considered safe due to strong boundary markers.
- Ingestion points: External MP4 URLs provided by the user and processed via
upload_media_from_urlin SKILL.md. - Boundary markers: The instructions explicitly mandate human review and confirmation in Step 5 ("Ask the user to confirm") and Step 7 ("After the user confirms") before any destructive or public-facing actions are taken.
- Capability inventory: Media upload (
upload_media_from_url), account management (list_accounts), and post creation/scheduling (create_post,schedule_post) as described in SKILL.md. - Sanitization: The skill verifies media metadata against platform-specific limits (length, size, aspect ratio) in Step 2.
Audit Metadata