bedrock-fine-tuning

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). Flagged because the skill passes S3 URIs at runtime (e.g., "s3://bucket/prompts.jsonl" and multiple "s3://my-bucket/..." examples) as required training/distillation datasets that directly supply prompts/completions which control the model's behavior.