Skills
skills/davila7/claude-code-templates/speculative-decoding/Gen Agent Trust Hub

speculative-decoding

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The file references/medusa.md provides instructions to download code from an untrusted GitHub repository: https://github.com/FasterDecoding/Medusa.
  • REMOTE_CODE_EXECUTION (MEDIUM): The file references/medusa.md contains a shell script block that instructs the user or agent to clone and immediately execute a training script (python medusa/train/train.py) from an untrusted source.
  • EXTERNAL_DOWNLOADS (LOW): The file references/lookahead.md references an untrusted external GitHub repository https://github.com/hao-ai-lab/LookaheadDecoding as a primary source for the algorithm implementation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 04:55 PM