claude-real-video-for-agents

Warn

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the user or agent to execute a script named install-skill.sh to automate the installation. The content of this script is not included in the analyzed file set, which constitutes a risk of executing unverified code.
  • [DATA_EXFILTRATION]: The tool includes specific command-line flags (--cookies and --cookies-from-browser) that allow the agent to access sensitive session data from local browser profiles (Chrome, Safari, Firefox, Edge) or cookie files. This is a mechanism for accessing credential-heavy data.
  • [PROMPT_INJECTION]: This skill is designed to ingest and process untrusted external data from video URLs and local files. This creates a surface for indirect prompt injection, where malicious instructions could be embedded in visual frames or transcriptions. The skill attempts to mitigate this with a textual warning instructing the agent to treat media content as data rather than instructions.
  • [COMMAND_EXECUTION]: The installation instructions include the use of sudo apt install ffmpeg, which involves acquiring elevated system privileges.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 12, 2026, 09:04 AM
Security Audit — agent-trust-hub — claude-real-video-for-agents