byted-kickart-video-analyzer
Fail
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill's upgrade process in
SKILL.mdinstructs the agent to execute a command string (install_command) received from a remote API response (upgrade.py). This allows a remote server to trigger arbitrary shell command execution on the host environment. - [REMOTE_CODE_EXECUTION]: The skill performs runtime installation of Python dependencies via
python3.12 -m pip install -r ./scripts/requirements.txt. Installing unpinned or unverified packages at runtime is a supply chain risk. - [CREDENTIALS_UNSAFE]: The skill instructions direct the agent to request that users provide
ACCESS_KEY_IDandSECRET_ACCESS_KEYdirectly in the chat session if they are not found in the environment. This practice exposes raw credentials in the chat logs and agent memory. - [EXTERNAL_DOWNLOADS]: The skill is designed to download content from arbitrary user-provided URLs using
curl -Linto the/tmpdirectory for analysis, creating a path for processing potentially malicious external data. - [COMMAND_EXECUTION]: The skill makes extensive use of subprocess execution to run Python scripts, download files, and manage environment variables (
export) based on dynamic user input and remote API data.
Recommendations
- AI detected serious security threats
Audit Metadata