Coordinates multi-level AI agent tasks with persistent repository-based memory and GitHub/Shortcut sync.
npx skills add https://github.com/stolinski/s-stack --skill dex-stolinski
Please help me install this Agent Skill. Skill: dex Source: https://github.com/stolinski/s-stack/tree/main/skills/dex Command: npx skills add https://github.com/stolinski/s-stack --skill dex-stolinski
Dex solves the challenge of coordinating multi-step work across sessions by storing tasks and decisions in a persistent memory system, enabling backlogs and long-term context preservation for AI agents.
Create an epic and break it into tasks and subtasks, then rely on Dex to persist context across sessions.
High-intent search queries and answers about installing and using this skill.
Persistent task memory for AI agents is maintained by storing tasks and decisions as repository files, ensuring context and progress survive beyond the current session for continuous coordination.
Yes, you can optionally sync your backlog with GitHub Issues. This synchronization creates a permanent record of your tasks and development workflow directly within your GitHub repositories.
You can organize complex multi-step work using a 3-level hierarchy consisting of Epics, Tasks, and Subtasks, allowing you to break down large projects while preserving context and dependencies.
The best way to preserve context for cross-session task handoffs is by capturing decisions, progress, and rationale in a persistent memory system that stores project files with optional GitHub or Shortcut sync.
Yes, repository file storage for task management supports dependencies and blocking. Tasks can block others, reflecting real-world workflows and ensuring complex multi-step work is coordinated accurately.
KarlinskySConfigure GitHub authentication via HTTPS tokens, SSH keys, or gh CLI login.
cstarleaTransform user requests into structured GitHub epics with child issues.
open-horizon-labsConvert session context into Jira subtasks under a specified parent.
himanshu231204Configure persistent self-directing agent workflows with Claude Code and MCP servers.
