
Query Cognee Memory with recall()
Teaches an agent to search cognee memory with recall(), pick search types, and debug empty results
- Repository
- topoteretes/cognee
- Stars
- 33,002
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This skill teaches an AI agent how to query cognee, an open-source memory system that stores information as a graph. It explains how to call recall(), which search strategy gets chosen automatically, how to narrow searches to specific datasets, node sets or sessions, and how to read the tagged results. It also covers common problems such as empty results, permission errors and the memory still warming up.
Great for
- Developers building an app or agent on top of cognee who need to query stored memory
- Engineers debugging why a cognee recall returns empty or unexpected results
- Teams deciding which search type or scope to use for a given question
Skip it if
- People who only want to store or ingest data into cognee, since this skill covers querying only
- Users without a cognee installation and a configured LLM key, because most search types need one and the skill assumes the library is set up
- Use with CYPHER queries on untrusted input: the skill notes that raw Cypher can write to the graph and is off by default
Try asking
- “Search my cognee memory for where Einstein was born and show the sources”
- “Use recall() restricted to the research dataset to answer a question about v2 changes”
- “Return only the context cognee would send to the LLM instead of a final answer”
- “Why does my cognee recall return an empty list for this shared dataset?”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
cognee-recall.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip cognee-recall.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip cognee-recall.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 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
- Pin query_type explicitly when you need predictable behaviour, since automatic routing can change the search type.
- Narrow searches with datasets or dataset_ids for speed, and remember that passing query_type drops session sources.
- If results are empty, check whether the graph is still building (memory_warming_up) or whether the user has permission to read the dataset.
What's inside
- 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.