스킬 deepchem 감사 이력
🧪

감사 이력

deepchem - 5 감사들

감사 버전 5

최신 낮은 위험

Jan 17, 2026, 12:26 AM

Legitimate molecular ML toolkit with standard ML library behaviors. Downloads pretrained models from HuggingFace and saves checkpoints locally. No suspicious network endpoints, credential access, or code execution patterns. Static findings are false positives from markdown documentation parsing.

7
스캔된 파일
2,639
분석된 줄 수
6
발견 사항
claude
감사자
중간 위험 문제 (1)
Pretrained model downloads from HuggingFace
The transfer_learning.py script downloads pretrained models from HuggingFace (line 61: model_id='seyonec/ChemBERTa-zinc-base-v1'). Standard ML library behavior for transfer learning workflows.
낮은 위험 문제 (1)
Model checkpoint storage to local directories
Scripts save model checkpoints to local directories (line 125: model_dir='./grover_pretrained'). Standard ML practice for model persistence.

감사 버전 4

낮은 위험

Jan 17, 2026, 12:26 AM

Legitimate molecular ML toolkit with standard ML library behaviors. Downloads pretrained models from HuggingFace and saves checkpoints locally. No suspicious network endpoints, credential access, or code execution patterns. Static findings are false positives from markdown documentation parsing.

7
스캔된 파일
2,639
분석된 줄 수
6
발견 사항
claude
감사자
중간 위험 문제 (1)
Pretrained model downloads from HuggingFace
The transfer_learning.py script downloads pretrained models from HuggingFace (line 61: model_id='seyonec/ChemBERTa-zinc-base-v1'). Standard ML library behavior for transfer learning workflows.
낮은 위험 문제 (1)
Model checkpoint storage to local directories
Scripts save model checkpoints to local directories (line 125: model_dir='./grover_pretrained'). Standard ML practice for model persistence.

감사 버전 3

낮은 위험

Jan 7, 2026, 12:59 AM

Legitimate molecular ML toolkit with standard ML library behaviors. Downloads pretrained models from HuggingFace and saves checkpoints locally. No suspicious network endpoints, credential access, or code execution patterns.

6
스캔된 파일
1,349
분석된 줄 수
6
발견 사항
claude
감사자
중간 위험 문제 (1)
Pretrained model downloads from HuggingFace
The transfer_learning.py script downloads pretrained models from HuggingFace (line 61: model_id='seyonec/ChemBERTa-zinc-base-v1'). This is standard ML library behavior for transfer learning workflows but represents external code execution from HuggingFace model hub.
낮은 위험 문제 (1)
Model checkpoint storage to local directories
Scripts save model checkpoints to local directories (line 125: model_dir='./grover_pretrained'). This is standard ML practice but could theoretically allow model file overwrites.

감사 버전 2

낮은 위험

Jan 7, 2026, 12:59 AM

Legitimate molecular ML toolkit with standard ML library behaviors. Downloads pretrained models from HuggingFace and saves checkpoints locally. No suspicious network endpoints, credential access, or code execution patterns.

6
스캔된 파일
1,349
분석된 줄 수
6
발견 사항
claude
감사자
중간 위험 문제 (1)
Pretrained model downloads from HuggingFace
The transfer_learning.py script downloads pretrained models from HuggingFace (line 61: model_id='seyonec/ChemBERTa-zinc-base-v1'). This is standard ML library behavior for transfer learning workflows but represents external code execution from HuggingFace model hub.
낮은 위험 문제 (1)
Model checkpoint storage to local directories
Scripts save model checkpoints to local directories (line 125: model_dir='./grover_pretrained'). This is standard ML practice but could theoretically allow model file overwrites.

감사 버전 1

낮은 위험

Jan 7, 2026, 12:59 AM

Legitimate molecular ML toolkit with standard ML library behaviors. Downloads pretrained models from HuggingFace and saves checkpoints locally. No suspicious network endpoints, credential access, or code execution patterns.

6
스캔된 파일
1,349
분석된 줄 수
6
발견 사항
claude
감사자
중간 위험 문제 (1)
Pretrained model downloads from HuggingFace
The transfer_learning.py script downloads pretrained models from HuggingFace (line 61: model_id='seyonec/ChemBERTa-zinc-base-v1'). This is standard ML library behavior for transfer learning workflows but represents external code execution from HuggingFace model hub.
낮은 위험 문제 (1)
Model checkpoint storage to local directories
Scripts save model checkpoints to local directories (line 125: model_dir='./grover_pretrained'). This is standard ML practice but could theoretically allow model file overwrites.