cortex

Build a hierarchical Cortex memory of any decomposable concept the agent can reach — a folder tree, a code repo, a website, a Notion or Linear workspace, a research topic, a document, an API surface, a person's body of work, a book, a course archive, anything that has a root and meaningful sub-parts. Trigger whenever the user asks to ingest, map, memorize, scan, survey, or build a complete picture of an existing thing — phrases like 'scan this', 'mappa Y', 'crea una memoria di X', 'fai una memoria del tema Z', 'ingest the OpenAPI spec into cortex', 'memorizza questo workspace Notion', 'costruisci una memoria di questo libro / repo / tema'. Spawns recursive exploration agents (one per node) that work over whatever medium reaches the target, and writes a tree of cortex notes mirroring the concept hierarchy. Always asks up-front how deep to go and previews the tree before writing anything.

Bootstrap Cortex on a fresh machine. Trigger this skill whenever the user wants to install, set up, configure, or first-time-use Cortex / the `cortex` CLI / the Cortex memory atlas — including phrases like 'install cortex', 'set up cortex', 'cortex non funziona', 'first time using cortex', 'configura cortex', 'help me get cortex working', or when the user says they cloned the repo and want to start. Also use this when `cortex` is missing, when `~/cortex/` doesn't exist, or when the user wants to point Cortex at a remote, switch image provider, or start the React atlas. Do NOT use this for daily memory work — that's the `cortex` skill.

Daily memory work in Cortex via the `cortex` CLI: capturing thoughts, writing structured notes, linking concepts, generating scene images, querying the knowledge graph. Use this skill whenever the user wants to remember, capture, save, jot down, log, link, ask, search, or work with their Cortex memory store / knowledge graph / atlas — including phrases like 'remember this', 'save a note about X', 'aggiungi una memoria', 'cerca nel cortex', 'cosa so di X', 'crea una scena per X', 'aggiorna l'immagine del nodo Y', or any reference to inbox / daily / notes inside `~/cortex/`. Trigger this whenever the user mentions cortex, memory, the atlas, scenes, hotspots, or wants to record/recall information about their projects or work. NEVER read files in `~/cortex/` directly with cat/Read — always use the `cortex` CLI.