Tessl • The management layer for AI agent skills

SkillPublic

managed-memory

Give an agent durable, cross-session long-term memory using Databricks MANAGED memory (the Unity Catalog memory-store REST APIs) as tools — governed by UC with no infra the customer needs to run. This works for either OpenAI Agents SDK or LangGraph templates. Use when: the agent should remember a user's (or a team/org's shared) preferences/facts/decisions across conversations; keywords 'long-term memory', 'managed memory', 'memory store', 'agentic memory'. This is separate from the self-hosted Lakebase memory solution with skills in (agent-openai-memory / agent-langgraph-memory).

Quality

79%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/managed-memory/SKILL.md

Quality

Content

63%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A detailed, highly actionable skill body with clear step sequencing and concrete executable code for the OpenAI path, weakened by notable repetition and a monolithic single-file structure with no progressive disclosure into reference files.

Suggestions

Deduplicate the 'semantic retrieval with BM25 keyword boosting' query guidance, which appears verbatim in the search_memory docstring, the Step-3 blockquote, and the Limits & errors section — keep one canonical copy and link to it.

Split the six-endpoint API reference comment, the full OpenAI tool implementations, and the Limits & errors table into reference files (e.g. references/api.md, references/tools_openai.md) and link from SKILL.md to reduce the monolithic body.

Provide complete copy-paste LangGraph tool definitions instead of the 'translation sketch' with 'Define save_memory ...' placeholders, so the LangGraph path is as actionable as the OpenAI one.

Dimension	Reasoning	Score
Conciseness	Mostly efficient and accurate, but the semantic-search/BM25 query guidance is repeated verbatim in three places (search_memory docstring, the Step-3 blockquote, and Limits & errors) and the scope/isolation discussion recurs across Concepts, Step 1, Step 4, the Scope strategy section, and Notes.	3 / 5
Actionability	The OpenAI path is fully copy-paste executable (complete REST curl calls, full Python tool implementations, wiring snippets), but the LangGraph variant is an explicit sketch with 'Define save_memory ...' placeholders rather than complete definitions.	4 / 5
Workflow Clarity	Steps 1–5 are clearly sequenced with a Test section and a Troubleshooting feedback table, plus fail-closed scope guards and a store-name confirmation check; minor validation gaps keep it just below a fully checkpointed 5.	4 / 5
Progressive Disclosure	Good in-page structure (headers, tables, step sequencing) but the entire ~600-line skill is one monolithic file with no bundle files — the API endpoint reference, full tool implementations, and Limits & errors reference that belong in separate files are all inlined.	3 / 5
	Total	14 / 20 Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that clearly states what the skill does and when to use it, with explicit trigger keywords and an explicit boundary against the related Lakebase memory skills. The only gap is that the six tools are not enumerated.

Dimension	Reasoning	Score
Specificity	Names the domain ('durable, cross-session long-term memory' via 'Unity Catalog memory-store REST APIs ... as tools') and concrete recall behavior ('remember ... preferences/facts/decisions'), but does not enumerate the six tools, leaving a minor coverage gap.	4 / 5
Completeness	Explicitly answers both what ('Give an agent durable, cross-session long-term memory using Databricks MANAGED memory ... as tools') and when ('Use when: the agent should remember ... across conversations; keywords ...') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Includes an explicit keyword list ('long-term memory', 'managed memory', 'memory store', 'agentic memory') that users would plausibly say, with good synonym coverage; a few more natural phrasings (e.g. 'remember things about me') are missing.	4 / 5
Distinctiveness Conflict Risk	Carves a clear niche and explicitly distinguishes itself from the self-hosted Lakebase skills ('separate from the self-hosted Lakebase memory solution with skills in agent-openai-memory / agent-langgraph-memory'), minimizing conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (624 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: databricks/app-templates
Path: .claude/skills/managed-memory/SKILL.md
Revision: 3826450

Reviewed: 26 days ago

Discovery Implementation Validation

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.

managed-memory

Table of Contents

Is this your skill?