pp-twelvelabs
Warn
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
twelvelabs-pp-clitool usingnpx -y @mvanhorn/printing-press-libraryorgo install github.com/mvanhorn/printing-press-library/library/ai/twelvelabs/cmd/twelvelabs-pp-cli@latest. These resources originate from third-party repositories. - [REMOTE_CODE_EXECUTION]: The installation process involves downloading and executing code from external npm and GitHub repositories to provide core skill functionality.
- [DATA_EXFILTRATION]: The CLI tool supports a
--deliver webhook:<url>flag, allowing the agent to send command outputs—potentially including video transcripts, summaries, and metadata—to an arbitrary external URL. - [COMMAND_EXECUTION]: The skill performs shell commands to interact with the local file system, such as reading local video files for uploading and using
ffmpegto write clip files to the disk. - [PROMPT_INJECTION]: The skill processes untrusted video analysis data, creating an indirect prompt injection surface. Ingestion points:
video-brief,summarize,analyze, andvideo-searchresults in SKILL.md. Boundary markers: Absent; command results are returned without delimiters. Capability inventory: The skill has file system access and network exfiltration capabilities via its own CLI flags. Sanitization: No sanitization of ingested video content or metadata is specified.
Audit Metadata