Primary-source register
What this course is built on.
Product-specific claims use current official documentation. Interfaces and feature names change, so this register should be reviewed quarterly.
OpenAI
- Desktop app
- Projects, files, tools, browser work, and long-running tasks in one desktop workspace.
- Codex best practices
- Context, task briefs, reusable guidance, configuration, testing, review, skills, and automation.
- Prompting
- Goal, context, output, boundaries, follow-up, and evidence-aware finishing checks.
- Permissions
- Approvals and boundaries for agent actions.
Anthropic
- Claude Code overview
- Agentic work across files, tools, commands, project guidance, skills, and multiple surfaces.
- Best practices
- Specific context, plan-first work, verification signals, reusable instructions, and scaling patterns.
- Security
- Permissions, sandboxing, working-directory boundaries, prompt-injection risk, and user responsibility.
- Custom Gems guidance
- Task, context, persona, output format, and uploaded knowledge for reusable chat assistants.
Editorial note
The course does not claim that a single AI product is universally best. It teaches a transferable operating model and names product-specific surfaces only where the primary documentation supports them.