# The Agent Behind the Curtain: Building the Oz Cloud Agent Platform — Safia Abdalla, Warp

## Executive summary

The presentation details the architecture and philosophy behind Warp's cloud agent platform, emphasizing that a successful developer tool must 'absorb complexity before it reaches the user.' The platform achieves this by providing structured environments (sandboxes) and flexible integration points (multi-harness support). Core functionality revolves around advanced agent orchestration—managing complex workflows across multiple sub-agents via both prompting (`/orchestrate`) and comprehensive API exposure. This approach allows non-developers to build sophisticated tooling on top of the system, significantly automating tasks like issue triage and pull request reviews in open-source repositories.

## Key takeaways

- Platform Complexity Absorption: A core principle is that platforms must hide underlying infrastructure complexity from the user. This includes managing where agents run (sandboxes) and supporting diverse developer preferences without fragmentation.
- Multi-Harness Support: The platform accommodates various developer workflows and tools (harnesses), ensuring consistency by providing structured guardrails around the chosen tool, regardless of whether it is a custom or native solution.
- Agent Orchestration: Since real engineering rarely fits into one prompt, agents are designed to orchestrate sub-agents for complex tasks. This can be triggered via a dedicated command (e.g., `/orchestrate`) or programmatically through the exposed API.
- API Composability: All key primitives—including spinning up agents, managing compute environments, and handling artifacts—are exposed via a robust API. This allows external users (even non-engineers) to build custom tooling (e.g., Slack bots) on top of the platform.

## Technical details

- Agent Execution Environment: Agents require isolated environments, or sandboxes, to run in the cloud. The platform supports both managed and self-hosted options, allowing teams to integrate agent workloads with their existing private infrastructure.
- Agent Workflow Automation: In open-source contributions (e.g., Warp's repo), agents automate the Software Development Life Cycle (SDLC) by triaging issues, researching codebases, drafting specifications, and performing agent-managed PR reviews, reserving human review for high-signal items.
- System Design Philosophy: The system is conceptualized not as a 'software factory,' but as a 'potter's workshop.' This metaphor emphasizes that the value lies in the repeatable, scalable system (the process) that supports human creativity and iteration.

## Practical implications

- Build engineers can leverage agent orchestration to automate complex, multi-step CI/CD processes that span multiple services or repositories.
- The emphasis on self-hosted sandboxes allows for the integration of AI agents into highly regulated or proprietary internal infrastructure without requiring cloud migration.
- By exposing primitives via a comprehensive API, teams can build specialized tooling (e.g., custom dashboards, monitoring bots) that interact with agentic workflows and artifacts.

## Topics

AI Agents, Cloud Computing, Developer Tooling, API Design, Workflow Automation, Open Source Development, Warp, Jupiter Notebook core team

Source: https://www.youtube.com/watch?v=L173Z8DpaJg