The Agent Skills Directory

[SAFE]: The primary content of this skill is educational documentation and Markdown-based templates. It provides patterns for agent orchestration, security guidelines, and best practices for developing reliable and secure AI agents.- [COMMAND_EXECUTION]: The provided utility scripts/scaffold_workflow.py performs standard file system operations such as creating directories and writing template files to initialize a project. It uses standard Python libraries and operates strictly within the scope of its documented purpose as a developer scaffolding tool.- [EXTERNAL_DOWNLOADS]: The documentation references resources and official guidance from trusted organizations including Microsoft, OpenAI, Anthropic, and IBM. These references are for documentation and research purposes and do not initiate unverified or dangerous code execution.- [PROMPT_INJECTION]: Prompt templates included in the skill provide strategies for managing agent behavior (such as forcing delegation). These are standard prompt engineering techniques for structured workflows and do not attempt to override platform-level safety features.- [DATA_EXFILTRATION]: No suspicious network activity or unauthorized data access patterns were identified. The skill explicitly includes security instructions that warn against insecure practices like hardcoding credentials in source code.

agentic-workflow-guide