CtrlK
BlogDocsLog inGet started
Tessl Logo

lakebase-setup

Configure Lakebase for agent memory storage. Use when: (1) Adding memory capabilities to the agent, (2) 'Failed to connect to Lakebase' errors, (3) Permission errors on checkpoint/store tables, (4) User says 'lakebase', 'memory setup', or 'add memory'.

80

3.40x
Quality

72%

Does it follow best practices?

Impact

92%

3.40x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./agent-langgraph-advanced/.claude/skills/lakebase-setup/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

100%

61%

Add Persistent Memory to a LangGraph Agent App

Provisioned Lakebase config for LangGraph agent

Criteria
Without context
With context

Memory extra dependency

100%

100%

Provisioned resource type

0%

100%

Correct instance name field

0%

100%

database_name field

0%

100%

CAN_CONNECT_AND_CREATE permission

100%

100%

LAKEBASE_INSTANCE_NAME via value_from

0%

100%

EMBEDDING_ENDPOINT in databricks.yml

0%

100%

EMBEDDING_DIMS in databricks.yml

0%

100%

.env LAKEBASE_INSTANCE_NAME

0%

100%

.env EMBEDDING settings

100%

100%

CLI profile included

100%

100%

bundle run after deploy

50%

100%

100%

66%

Wire Up Autoscaling Lakebase for an OpenAI SDK Agent

Autoscaling Lakebase config for OpenAI agent

Criteria
Without context
With context

Autoscaling resource type

100%

100%

Branch field full path

100%

100%

Database field full path with ID

0%

100%

CAN_CONNECT_AND_CREATE permission

100%

100%

LAKEBASE_AUTOSCALING_PROJECT env var

0%

100%

LAKEBASE_AUTOSCALING_BRANCH env var

0%

100%

EMBEDDING_ENDPOINT in databricks.yml

0%

100%

EMBEDDING_DIMS in databricks.yml

0%

100%

.env autoscaling vars

0%

100%

.env EMBEDDING settings

0%

100%

Discovery commands use postgres API

0%

100%

Discovery commands include profile

100%

100%

78%

69%

Debug and Fix Lakebase Permission Errors on First App Start

SP permission grant workflow after deploy

Criteria
Without context
With context

Gets SP client_id via CLI

33%

100%

Runs grant script

0%

100%

Correct memory-type flag

0%

100%

Instance name in grant script

0%

100%

Profile on grant script

0%

100%

Bundle run command

0%

0%

Profile on bundle run

0%

0%

SP client_id used in grant

0%

100%

No deploy command

100%

100%

Repository
databricks/app-templates
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.