Skip to content
Get packs

Cognee Memory Skill

Lets an agent store documents and chat context in a knowledge graph and query it later

Source Download skill
Repository
topoteretes/cognee
Stars
33,002
Licence
Apache-2.0
Files
15 files

What it does

Once installed, the agent knows how to use the Cognee Python library as a long-term memory layer: it can ingest text, files, URLs or repositories with remember(), build a knowledge graph from them, and answer questions later with recall(), choosing a search strategy automatically. It also guides the agent on adding feedback loops, session memory, tagging with node_set, custom extraction models, and deleting data with forget(). The skill is a playbook of API guidance; the actual memory work is done by the Cognee library the user must install and configure.

Great for

  • Python developers building agents that need memory across sessions
  • Teams turning internal documents into a searchable knowledge graph
  • Data engineers who want to explore graph-based RAG with Cognee

Skip it if

  • Non-developers wanting a ready-made app, since this only guides an agent that writes Cognee Python code
  • Sensitive data without review: ingested content is sent to LLM and database services configured by the user, and remember() may send files and URLs onward
  • Projects not using Cognee, since the skill is specific to its API

Try asking

  • “Store these meeting notes in a Cognee dataset called research and then answer what decisions were made in Q1”
  • “Build a knowledge graph from the PDFs in ./papers and show which entities are most connected”
  • “Set up session-based memory so my support agent remembers each customer's preferences”
  • “Delete the old tutorial dataset from my Cognee memory without touching other datasets”
Step-by-step guide: how to use Cognee Memory SkillInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 15 original files, the Apache-2.0 licence and an attribution note.

    cognee.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip cognee.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip cognee.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Start with remember() and recall() and only use advanced options like graph_model or custom pipelines when needed
  • Use dry_run=True in remember() to estimate LLM token cost before ingesting large sources
  • Tag data with node_set and dataset_name to keep per-user or per-project memory separated, and avoid prune_system outside test teardown since it drops users and ACLs

What's inside

  • __init__.py
  • __main__.py
  • alembic/__init__.py
  • alembic/env.py
  • alembic/frozen_schema.py
  • alembic/versions/18f01b0a0b4c_create_principal_configuration_table.py
  • alembic/versions/1a58b986e6e1_enable_delete_for_old_tutorial_notebooks.py
  • alembic/versions/1c22e6cb5aec_reconcile_to_frozen_schema.py
  • alembic/versions/1d0bb7fede17_add_pipeline_run_status.py
  • alembic/versions/1daae0df1866_incremental_loading.py
  • alembic/versions/211ab850ef3d_add_sync_operations_table.py
  • alembic/versions/24f5d4f64d0d_add_session_lifecycle_tables.py
  • alembic/versions/45957f0a9849_add_notebook_table.py
  • alembic/versions/46a6ce2bd2b2_expand_dataset_database_with_json_.py
  • skill.md

Redistributed unmodified from topoteretes/cognee under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.