core

Initialize and maintain Mycelium living repositories with structured project memory.

68|2|Updated Feb 25, 2026
One-click install
npx skills add https://github.com/arjunrajlaboratory/mycelium --skill core-arjunrajlaboratory
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: core
Source: https://github.com/arjunrajlaboratory/mycelium/tree/main/skills/core
Command: npx skills add https://github.com/arjunrajlaboratory/mycelium --skill core-arjunrajlaboratory

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Analytical projects lose hard-won knowledge: decisions, gotchas, and conventions disappear between sessions and team members. This Skill scaffolds and maintains a Mycelium living repository, giving every project a persistent .living/ memory layer that records decisions, learnings, findings, and conventions across sessions. ## Core Features & Use Cases - Repository Initialization & Migration: Scaffold new or existing repos into mycelium-enabled living repositories with manifests, convention packs, hooks, and a .living/ knowledge layer; migrate older Mycelium repos to the current spec. - Knowledge Crystallization: Extract recurring patterns from .living/learnings.md into conventions, crystallize scientific findings into topic-based files with registries, and promote transferable knowledge to a global ~/.mycelium/knowledge/ system. - Cross-Project Transfer & Contribution: Cross-pollinate learnings across sibling projects in a meta-project, and package generalized conventions for contribution back to the network. - Use Case: A bioinformatics lab initializes Mycelium on an RNA-seq project; after each analysis session, decisions and learnings are logged, recurring pitfalls crystallize into conventions, and relevant insights automatically transfer to the lab's image-analysis project. ## Quick Start Ask the assistant to set up Mycelium and initialize this repository as a living repo.

Frequently Asked Questions about core

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I initialize a Mycelium living repository?

Ask the assistant to set up Mycelium or initialize a living repo. The init mode runs init_repo.py to scaffold the .living/ directory, manifests, convention packs, hooks, and guidance files, then validates the structure with validate_structure.py.

How do I crystallize learnings into conventions?

Trigger crystallize mode by asking to review learnings or extract patterns. The skill scans .living/learnings.md for 3+ related entries, optionally runs detect_recurrence.py to find near-duplicate clusters, and drafts new conventions in .living/generated-conventions/.

Does Mycelium work with both Claude Code and Codex?

Yes, Mycelium supports both hosts. Claude Code uses repository-local hooks registered in .claude/settings.local.json, while Codex uses five plugin-bundled hooks that must be trusted via the /hooks command in the Codex CLI.

How does cross-project knowledge transfer work in Mycelium?

Transfer mode requires a meta-project layout where a parent directory with .living/ contains sibling projects. A background subagent scans recent learnings across siblings and auto-applies relevant ones to each project's learnings.md with audit reports.

What are the limitations of Mycelium's transfer and contribute modes?

Transfer mode is dormant unless projects are nested under a meta-project with its own .living/ directory; plain sibling repos do not qualify. Contribute mode is manual only and requires a previously crystallized generated convention to package.