configuration-backends
Configure Cognee runtime, optional extras, provider settings, storage/database backends, paths, environment variables, and backend troubleshooting.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- configuration-backends
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Configure Cognee runtime, optional extras, provider settings, storage/database backends, paths, environment variables, and backend troubleshooting.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/backend-matrix.md4.5 KB
- 参考文档references/configuration.md13.5 KB
- 参考文档references/troubleshooting.md3.2 KB
- 脚本scripts/check_cognee_environment.py3.8 KB
03 SKILL.md 正文
Cognee Configuration and Backend Router
Use this sub-skill when the task is about installing Cognee, choosing optional
extras, setting environment variables, configuring cognee.config, selecting
LLM/embedding providers, changing storage paths, or diagnosing database/cache
backend setup.
Route here for
- Python/package install requirements and optional extras selection.
.envstyle setup for LLMs, embeddings, storage, databases, cache, tracing, auth posture, and local model providers.- Runtime setters under
cognee.config, including LLM, embedding, graph, vector, relational, migration, data-root, and system-root configuration. - Backend choices: default SQLite/LanceDB/Ladybug/Kuzu, Postgres/pgvector, Neo4j, Turso/libSQL, Neptune Analytics, S3 storage, Redis/session cache, and local LLM/embedding providers.
- Safe preflight checks that should not call LLMs, connect to databases, print secrets, or mutate data.
Route away
- Running memory ingestion/search/improvement flows: read core-memory.
- CLI invocation, long-running API servers, Docker, MCP, UI, CORS deployment, and service hardening: read api-cli-services.
- Custom graph schemas, pipelines, ontology resolvers, migrations, and advanced graph operations: read advanced-graphs-pipelines.
Operating workflow
-
Read configuration.md for install range, environment tiers, runtime config classes, and safe
cognee.configsetters. -
Read backend-matrix.md before changing extras or backend provider names. Treat every external service and credentialed provider as optional unless the user explicitly selected it.
-
If the user reports an error, read troubleshooting.md and identify whether the failure is missing credentials, missing optional extras, provider inference, vector dimensions, access-control/handler mismatch, cache/S3 path handling, or database subprocess/file locking.
-
For a non-mutating local preflight, run the bundled checker after Cognee is installed:
python scripts/check_cognee_environment.py --help python scripts/check_cognee_environment.py --jsonThe checker redacts secrets and absolute paths, constructs config objects safely, and reports optional modules as present/missing.
Safety rules
- Never ask the user to paste secret values. Ask whether a secret is set, or use placeholders in examples.
- Do not install broad extras by default. Select the smallest extra set needed for the chosen provider/backend.
- Set environment variables before importing
cogneein service or notebook entrypoints when those variables must affect pydantic settings at import time. - Do not use this sub-skill to execute real memory flows or service startup; it owns configuration only.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).