The Agent Skills Directory

[SAFE]: The skill's primary function is to enforce rigorous scientific methodology and provide a diagnostic framework for deep learning research. The instructions focus on improving the quality of research decisions and debugging processes.
[SAFE]: The silent-bugs.md file is a high-quality reference document cataloging 191 common implementation errors in various deep learning architectures. It includes code snippets for detection and mitigation, along with links to authoritative sources such as official GitHub issues (PyTorch, HuggingFace, DeepSpeed) and peer-reviewed research papers (arXiv).
[SAFE]: No malicious patterns, such as prompt injection, unauthorized data access, credential exfiltration, or remote code execution, were found in any of the skill's files. The skill correctly utilizes the agent environment for parallel task execution and knowledge management.
[SAFE]: The metadata and documentation are consistent with the skill's stated purpose, and no obfuscation or hidden content was detected.

research-collaborator