speculative-decoding

Warn

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to clone and install code from external, unverified GitHub repositories: 'https://github.com/FasterDecoding/Medusa' and 'https://github.com/hao-ai-lab/LookaheadDecoding'. Downloading and executing code from third-party sources without integrity checks or version pinning can expose the environment to malicious code if the repositories are compromised.
  • [COMMAND_EXECUTION]: The installation guide includes shell commands for cloning repositories and performing editable installs ('pip install -e .'). This process involves executing setup scripts (e.g., 'setup.py' or 'pyproject.toml' logic) from the downloaded source code, which can perform arbitrary actions on the host system.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 28, 2026, 06:07 PM