Mymir FAQs

Question 1

How does Mymir integrate with existing agent workflows, particularly with Claude Code?

Accepted Answer

Mymir ships as a Claude Code plugin, bundling an MCP server, specialized agents (brainstorm, decompose, manage), and a '/mymir' skill. This allows for natural language interaction, auto-invoking Mymir's capabilities when discussing projects or tasks.

Question 2

What are the core concepts behind Mymir's functionality?

Accepted Answer

Mymir is built on two core concepts: the 'Context network' – a living map of project decisions, relationships, and abandoned paths – and the 'Context retrieval interface,' which allows agents to query and utilize this knowledge precisely when needed.

Question 3

What are the key differences between Structure mode and Graph mode in Mymir's UI?

Accepted Answer

Structure mode in Mymir's web UI focuses on task management with a task list and detail panel for refining specs and tracking progress. Graph mode overlays the context network, allowing users to visualize tasks, decisions, and dependencies, offering a broader view of project relationships.

Question 4

Which stages of the project lifecycle does Mymir support?

Accepted Answer

Mymir manages the full project lifecycle, supporting stages from Brainstorm, Decompose, and Refine to Plan, Execute, and Track. It provides stage-appropriate context to agents automatically at each step.

Question 5

What is Mymir and how does it help engineering teams?

Accepted Answer

Mymir is an agent-native project context management tool designed for engineering. It creates a live knowledge base that guides AI agents through the entire engineering lifecycle, eliminating the need to constantly re-brief them and ensuring consistent, informed project execution.

Mymir

Mymir

Key Features

Use Cases

Key Features

Use Cases