It feels like we have a solid foundation for most typical AI workflows:
- Choose a model and an agentic harness which are appropriate for the sort of work you need to complete in a project. and for the environment/requirements surrounding a project.
- Use /goal or have your agent manually build a plan that can be executed over multiple long running sessions.
- Focus on building detailed engineering requirements, and comprehensive tests for any workflow in which the result can be comprehensively validated
- Orchestrate model usage so that larger, more expensive models build plans and complete complex reasoning steps which require lots of world knowledge, delegate easier tasks to faster, cheaper, small models, and check the work of smaller models.
- Manage context length by following a concrete plan which produces artifacts that can be inspected and used across steps. Your plan and artifacts should enable the long running process to continue at any point Spawn sub-agents to complete task
- After you've completed long horizon tasks, build skill packages to encapsulate all the in-context learning needed to repeat steps in any process which will be executed again in the future. Don't waste the tokens going through the same trial and error process again.
- Along with skills, build agent.md files and extensions for your agent, to establish workflow guidelines, to remember learned patterns, and to extend an agent's capability