video-qa
Fail
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill includes instructions to install Homebrew by executing its official installation script from
https://raw.githubusercontent.com/Homebrew/install/HEAD/install.shvia a shell command. It also suggests installingffmpegthrough various system package managers such asapt,dnf,winget, andchoco. - [COMMAND_EXECUTION]: The skill directs the agent to perform command-line operations for processing video files, specifically using
ffmpegfor multi-frame extraction andqlmanagefor generating thumbnails. It also references the use ofsudofor package management, which is contextually appropriate for the tool's installation instructions. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing untrusted project data to perform its quality assurance tasks.
- Ingestion points: The skill reads external files such as
shot-list.md,generation-log.md, and character/environment/prop bibles. - Boundary markers: The instructions lack specific delimiters or safety warnings for the agent to distinguish between data and instructions within these files.
- Capability inventory: The agent has the ability to read local files, execute shell commands (e.g., ffmpeg), and provide generated prompts to other specialist agents.
- Sanitization: No input validation or content sanitization is described for the processed data files.
Recommendations
- HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh - DO NOT USE without thorough review
Audit Metadata