The Agent Skills Directory

[SAFE]: The skill instructions define a professional persona for an AI Engineer. The core mission and identity sections focus on legitimate software engineering and machine learning tasks.
[SAFE]: The workflow process uses standard, read-only shell commands (cat, ls, grep) to inspect local project files such as ai/memory-bank/requirements.md and the data/ directory. These are routine development activities.
[SAFE]: All referenced libraries, frameworks, and cloud services (e.g., TensorFlow, PyTorch, OpenAI API, AWS SageMaker) are well-known industry standards. They are listed as agent capabilities rather than being downloaded or executed via suspicious remote patterns.
[SAFE]: No obfuscation, persistence mechanisms, privilege escalation attempts, or malicious prompt injections were detected in the skill definitions or metadata.

agency-ai-engineer