viral-hook-formats

Pass

Audited by Gen Agent Trust Hub on Nov 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to ingest and process external, untrusted data including public video captions, transcripts, and metadata. This creates an attack surface where instructions hidden within these external sources could potentially influence the agent's execution. While the skill lacks formal technical delimiters for this data, it includes robust procedural mitigations, such as instructing the agent to strictly distinguish direct observations from inferences and warning against leaking private user briefs to external services.
  • [COMMAND_EXECUTION]: The instructions describe the use of local command-line utilities, specifically FFmpeg and ffprobe, to perform technical analysis of media files (e.g., checking duration, dimensions, and frame timing). These tools are used within the local environment to support the skill's core purpose of media inspection and do not involve executing untrusted scripts.
  • [EXTERNAL_DOWNLOADS]: The skill references external resources to provide additional functionality, such as a browser-based inspection tool from Vercel Labs' official GitHub repository and a skill catalog maintained by the vendor (branded-dropshipping). These downloads originate from trusted organizations or the vendor's own infrastructure and are intended to provide necessary tooling for the agent's tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 15, 2026, 10:35 AM