# Full Workshop: Setting Yourself Up for Success —Jason Liu, OpenAI Codex

## Executive summary

Jason Liu provides a comprehensive workshop on leveraging Codex and the Agents SDK for advanced knowledge work automation. The focus shifts from simple text generation to using AI agents (threads) as collaborative teammates capable of managing complex projects, interacting with external applications via 'Computer Use,' and maintaining long-running processes through automated 'heartbeats.' Key strategies include building custom skills/plugins, utilizing a personal memory vault, and treating the computer itself as an integrated component of the workflow.

## Key takeaways

- Systematic Context Management via Appshots: The 'Appshot' feature is highlighted as critical for providing deep context to AI models. Unlike simple screenshots, Appshots capture the entire accessibility tree of an application (e.g., Slack), allowing agents to perform complex actions with single function calls and greatly improving accuracy.
- Advanced Project Management with Threads: The concept of 'compaction' allows threads (pinned conversations) to maintain long-term context, acting like dedicated project managers. These threads can delegate tasks to sub-agents and communicate with each other, forming a collaborative team structure.
- Automating Long-Running Workflows (Heartbeats): Instead of manual checks, 'heartbeat' automations can wake up threads over time to perform continuous maintenance tasks (e.g., checking PR status, monitoring support issues), allowing for sustained project oversight.
- Extending AI Capabilities with Skills and Plugins: Skills are simple constructs (files/scripts) that solve specific problems (e.g., 'Review my code like Charlie'). Plugins connect the system to external services (Slack, Gmail), allowing agents to perform actions across different platforms.

## Technical details

- Agent Architecture & Workflow: The system moves from an IC enabled by an IDE to a 'team' of pinned threads, where the user acts as the manager. This structure allows for complex orchestration and delegation.
- Code Quality & Review: Skills can be built to automate code review processes (e.g., 'Review my code like Charlie'), analyzing past pull request feedback to maintain style guide adherence and find errors.
- Goal Setting & Iteration: The `slash goal` command defines a verification step that the agent must continuously check. The 'Ultra Goal' skill allows defining and editing goals within a file, enabling iterative project scope changes.
- System Integration & Control: The 'Computer Use' feature is emphasized as uniquely powerful because it can control any application (Slack, trading software) behind the scenes, unlike the Chrome extension which is limited to the browser.

## Practical implications

- Automating complex software development tasks like running continuous integration (CI) checks and managing pull request feedback loops.
- Improving developer productivity by using AI agents to handle repetitive context gathering, documentation, and cross-platform communication.
- Implementing 'Chief of Staff' threads that act as a single source of truth for project status across multiple disconnected systems (Slack, email, Git).
- Using the system to manage personal or professional operational tasks, such as booking travel or filing forms, by controlling native applications.

## Topics

AI Agents, Knowledge Work Automation, Software Development Lifecycle (SDLC), Computer Vision/Interaction, Plugin Ecosystems, Project Management, Continuous Integration (CI), Personal Monorepo Template, Codex App / Agents SDK

Source: https://www.youtube.com/watch?v=il1c1a2FufU