mem0-test-integration

Verifies Mem0 integrations by running native tests, smoke flows, and E2E checks.

Updated Jun 29, 2026
One-click install
npx skills add https://github.com/nhatnguyen1122/Agent-Memory-Eval --skill mem0-test-integration-nhatnguyen1122
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mem0-test-integration
Source: https://github.com/nhatnguyen1122/Agent-Memory-Eval/tree/main/mem0/skills/mem0-test-integration
Command: npx skills add https://github.com/nhatnguyen1122/Agent-Memory-Eval --skill mem0-test-integration-nhatnguyen1122

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After generating a Mem0 memory integration with /mem0-integrate, you need proof that it actually works — that dependencies install, existing tests still pass, new tests pass, and memory really flows end-to-end against a live API key. This Skill automates that verification and produces a pass/fail scorecard. ## Core Features & Use Cases - Precondition gating: Refuses to run unless .mem0-integration/ artifacts exist, the branch starts with mem0-integrate/, the working tree is clean, and the correct API key (MEM0_API_KEY or OPENAI_API_KEY) is present. - Two-pass test execution: Runs the repo's native suite with the feature flag unset (must behave like main) and set (new tests run), catching non-invasiveness violations as hard fails. - Real smoke and E2E tests: Exercises a live add/search/delete flow against Mem0 Platform or OSS with a disposable user_id, then runs the app via the plan's E2E recipe and asserts memory appears in user-visible output. - Use Case: After running /mem0-integrate on a feature branch, invoke this Skill to get a scorecard.json with per-check pass/fail, durations, log tails, and reproduction commands before human review. ## Quick Start Ask your assistant to run /mem0-test-integration on the current branch to verify the Mem0 integration and produce a scorecard.

Frequently Asked Questions about mem0-test-integration

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify a Mem0 integration after running /mem0-integrate?

Invoke /mem0-test-integration on the same branch and workspace. It installs dependencies, runs static checks, executes the native test suite in two passes, runs a live smoke test, and writes a scorecard to .mem0-integration/scorecard.md and scorecard.json.

What preconditions does mem0-test-integration require before running?

The .mem0-integration/ directory must exist with readable product.json, goal.md, and plan.md, the branch must start with mem0-integrate/, the working tree must be clean, and the same API key used during integration must be exported. Any failure exits with a written rationale.

Does mem0-test-integration work with both Mem0 Platform and OSS?

Yes. It reads the product field from product.json and uses MEM0_API_KEY with MemoryClient for Platform, or OPENAI_API_KEY with Memory and a local Qdrant store for OSS. If the repo ships a docker-compose Qdrant service, it starts and tears it down automatically.

Can I run mem0-test-integration in CI without prompts?

Yes. Set MEM0_TEST_CI=1 or pass --ci for non-interactive mode where keys must be in the environment and any failure exits non-zero. Flags like --skip-smoke, --skip-e2e, --only-smoke, and --only-e2e control which stages run.

What kinds of bugs does mem0-test-integration not catch?

It only catches compile and runtime bugs by design. Logical errors such as wrong data stored, incorrect user_id scoping, or retrieval at the wrong moment are explicitly out of scope and require human review, as noted in the scorecard's NOT checked section.

Why did mem0-test-integration exit with code 7?

Exit code 7 means a non-invasiveness violation: pre-existing tests failed with the feature flag unset, so the branch no longer behaves like main. This is a hard fail that the integrator's self-heal loop refuses to patch, and it must be surfaced to the user.