// Use when Codex should remember, recall, search, improve, or forget information using the Cognee CLI.
Use when ingesting, cognifying, or querying a codebase with Cognee CLI from Codex.
Use when launching, checking, or reporting on the local Cognee UI/backend through cognee-cli -ui.
Use when setting up, checking, or connecting Cognee through the cognee CLI from Codex.
Store data permanently in the Cognee knowledge graph. Accepts a data category (user, project, or agent) to tag the data with the correct node_set for filtered retrieval.
Search Cognee memory. Session memory is automatically searched on every prompt via hooks. Use this skill explicitly for permanent knowledge graph search, filtered category search, or when you need more results than the automatic lookup provides.
Sync session cache entries into the permanent Cognee knowledge graph. Run this to make session memory searchable, or it runs automatically at session end.
| name | memory |
| description | Use when Codex should remember, recall, search, improve, or forget information using the Cognee CLI. |
Use this skill when the user asks Codex to use Cognee as memory, add facts or documents, search a knowledge graph, recall prior context, or improve existing memory.
uv run cognee-cli ...; do not use MCP.remember for one-step ingestion and graph construction.add plus cognify when ingestion and processing should be staged.-d or --dataset-name; ask only if the dataset boundary is genuinely ambiguous..env files, private keys, token dumps, or unrelated generated artifacts.forget, delete, or --everything, get explicit user confirmation.One-step ingestion:
uv run cognee-cli remember <text-or-path> -d <dataset-name>
For larger or staged work:
uv run cognee-cli add <text-or-path> -d <dataset-name>
uv run cognee-cli cognify -d <dataset-name>
For long processing:
uv run cognee-cli remember <text-or-path> -d <dataset-name> --background
uv run cognee-cli cognify -d <dataset-name> --background
Memory-oriented recall:
uv run cognee-cli recall "<question>" -d <dataset-name> -f pretty
Search modes:
uv run cognee-cli search "<question>" -d <dataset-name> -t GRAPH_COMPLETION -f pretty
uv run cognee-cli search "<exact passage or citation need>" -d <dataset-name> -t CHUNKS -k 10 -f pretty
uv run cognee-cli search "<code question>" -d <dataset-name> -t CODE -k 10 -f pretty
Use -f json when downstream parsing is needed.
Enrich an existing graph:
uv run cognee-cli improve -d <dataset-name>
Bridge session feedback or Q&A into the graph:
uv run cognee-cli improve -d <dataset-name> -s <session-id>
For targeted enrichment:
uv run cognee-cli improve -d <dataset-name> --node-name <entity-name>
Use the narrowest deletion command possible and confirm first:
uv run cognee-cli forget --dataset <dataset-name>
uv run cognee-cli forget --dataset <dataset-name> --data-id <data-uuid>
Avoid uv run cognee-cli forget --everything unless the user explicitly asks
to delete all Cognee data.