Trusted — Risk Score 0/100
Last scan:1 day ago Rescan
0 /100
paper-to-pipeline
根据机器学习/深度学习论文的实验规划文档自动生成完整的 Python 实验 pipeline。支持数据预处理、模型构建、训练循环、评估指标、结果可视化。
A clean ML pipeline code generator with no security issues; reads experiment plans and writes standard Python ML code to a user-specified directory using only standard libraries.
Skill Namepaper-to-pipeline
Duration26.0s
Enginepi
Safe to install
No action needed. The skill is safe to use as described.
ResourceDeclaredInferredStatusEvidence
Filesystem READ READ ✓ Aligned scripts/generate_pipeline.py reads plan files via open()
Filesystem WRITE WRITE ✓ Aligned SKILL.md declares output to generated_experiment/ directory; code writes to user…
Network NONE NONE No network calls in any script
Shell NONE NONE No subprocess/os.system/os.popen calls found
Environment NONE NONE No os.environ access in generate_pipeline.py
Skill Invoke NONE NONE No skill invocation patterns detected
Clipboard NONE NONE No clipboard access
Browser NONE NONE No browser/web automation
Database NONE NONE No DB connections

File Tree

5 files · 52.0 KB · 1951 lines
Python 2f · 1224L Markdown 3f · 727L
├─ 📁 assets
│ └─ 📁 templates
│ └─ 🐍 image_classification.py Python 181L · 4.6 KB
├─ 📁 references
│ ├─ 📝 data-format.md Markdown 236L · 4.4 KB
│ └─ 📝 model-templates.md Markdown 342L · 9.8 KB
├─ 📁 scripts
│ └─ 🐍 generate_pipeline.py Python 1043L · 29.1 KB
└─ 📝 SKILL.md Markdown 149L · 4.0 KB

Security Positives

✓ No shell execution (subprocess, os.system, curl|wget, or pip install commands)
✓ No network requests — all imports are local standard library (argparse, os, re, yaml, pathlib, datetime)
✓ No credential harvesting — no access to ~/.ssh, ~/.aws, .env, or environment variable enumeration
✓ No obfuscation — no base64, no eval(atob()), no encoded payloads
✓ No persistence mechanisms — no cron jobs, startup scripts, or backdoor installation
✓ No prompt injection — generated code contains only standard ML boilerplate
✓ No sensitive file paths accessed
✓ Doc-to-code match is excellent — SKILL.md accurately describes all behavior
✓ All file writes are to user-specified output directories, declared in SKILL.md
✓ Dependencies are standard ML libraries (torch, torchvision, transformers) only in generated requirements.txt, not in the skill itself