moda-video-footage

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation includes the moda rushes install command, which downloads vendor-specific inference models and toolsets required for local video analysis.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and transcribes untrusted local video footage. * Ingestion points: Local footage is ingested via the moda rushes ingest command as described in references/footage.md. * Boundary markers: Explicit safety instructions are provided in references/reading-and-verifying.md, directing the agent to ignore instructions embedded in canvas data. * Capability inventory: The agent can extract files locally, upload to the cloud, and modify the project timeline using vendor-specific CLI tools. * Sanitization: The skill relies on behavioral instructions for the agent to distinguish data from tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 11:12 PM
Security Audit — agent-trust-hub — moda-video-footage