agentleFS
MicrosoftKnown publishermicrosoft/rayfin/.github/skills/run-cli-e2e-locally/SKILL.md

run-cli-e2e-locally

Skill646 starsChanged 24 days agoMITView on GitHub

Run the CLI E2E tests locally using your own user credentials instead of a service principal. Use when: run e2e locally, run cli tests, test with my credentials, local e2e, e2e user auth, run fabric tests locally.

  • Reads credentials

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Create packages/tools/cli-e2e/.env (this file is gitignored):

    line 25
  2. Distinctive:

    If the user doesn't know their tenant ID, suggest:

    line 40
  3. Very specific:

    If no valid user session exists, it automatically runs rayfin login --tenant <E2E_TENANT_ID> which opens the browser for interactive …

    line 58
  4. Distinctive:

    After authenticating once, subsequent test runs reuse the cached session without prompting again.

    line 59
  5. Very specific:

    When E2E_AUTH_MODE=user, it checks rayfin login status output.

    line 113
  6. Very specific:

    If not signed in as a user, it triggers rayfin login --tenant <E2E_TENANT_ID> (opens browser).

    line 114

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Help a developer run the CLI E2E tests locally using their own Azure AD credentials.

Prerequisites:

  • The repo is cloned and Rush dependencies are installed (rush update)
  • The CLI and E2E packages are built (rush build --to rayfin-cli-e2e)
  • The user has an Azure AD account with access to the target Fabric workspace

Steps

1. Create the local environment file

Create packages/tools/cli-e2e/.env (this file is gitignored):

E2E_AUTH_MODE=user
E2E_ENVIRONMENT=daily
E2E_TENANT_ID=<your-tenant-id>
E2E_CLI_SOURCE=source
USE_LOCAL_PACKAGES=true
RAYFIN_E2E_WORKSPACE_NAME=DO_NOT_DELETE_AppDev_E2E_Workspace
RAYFIN_TELEMETRY_OPTOUT=1
RAYFIN_ENCRYPTION_FALLBACK_ENABLED=true

Replace <your-tenant-id> with the user's Azure AD tenant ID.

If the user doesn't know their tenant ID, suggest:

rayfin login status

or check the Azure Portal under Microsoft Entra ID → Overview → Tenant ID.

2. Run the tests

From packages/tools/cli-e2e:

rushx test:e2e

Login is automatic. The test suite includes a globalSetup that checks for an existing user session before tests run. If no valid user session exists, it automatically runs rayfin login --tenant <E2E_TENANT_ID> which opens the browser for interactive sign-in. After authenticating once, subsequent test runs reuse the cached session without prompting again.

To run a specific test file:

rushx test:e2e -- src/init/init.test.ts

To run with coverage:

rushx test:e2e:coverage

To run in watch mode during development:

rushx test:e2e:watch

3. Run the Playwright browser E2E tests

First install Playwright browsers (one-time):

npx playwright install chromium

Then run the browser tests:

rushx test:e2e:browser

Note: The browser tests deploy to a Fabric workspace and navigate to the hosted URL. They require the user's account to have Contributor access to the DO_NOT_DELETE_AppDev_E2E_Workspace workspace. The Playwright tests do not use the Vitest globalSetup, so the user must have a valid session from a prior rushx test:e2e run or manual rayfin login --tenant <id>.

4. Troubleshooting

SymptomFix
E2E_AUTH_MODE=user requires E2E_TENANT_ID to be setFill in E2E_TENANT_ID in your .env file
Interactive login failed + browser did not openRun manually: rayfin login --tenant <id>
Silent token acquisition failedToken expired — re-run rayfin login --tenant <id> or delete ~/.rayfin/auth.json and retry
Workspace not foundVerify your account has access to DO_NOT_DELETE_AppDev_E2E_Workspace in the daily Fabric environment
Tests skip with Skipped: authMode === 'user'Expected — SP-specific tests are skipped in user mode
Tests skip with data-client tests require SP credentialsExpected — data-client tests need SP secrets
Keychain errors (WSL/containers)Set RAYFIN_ENCRYPTION_FALLBACK_ENABLED=true in .env
Windows libuv assertion crash during loginHarmless — the globalSetup checks CLI output content, not exit code

How It Works

  • Automatic login via globalSetup: The Vitest config references src/global-setup.ts which runs before all tests. When E2E_AUTH_MODE=user, it checks rayfin login status output. If not signed in as a user, it triggers rayfin login --tenant <E2E_TENANT_ID> (opens browser). On subsequent runs, the cached session is reused and login is skipped.
  • Output-based detection: On Windows, rayfin login may crash during process cleanup (libuv assertion) even after success. The setup checks for "Signed in successfully" in stdout rather than relying on exit code.
  • Auth state sharing: CLI commands in tests inherit the user's auth from ~/.rayfin/auth.json. Direct Fabric API calls acquire a token silently from the MSAL cache via the CLI's getAuthenticatedToken() function.
  • Test skipping: Service principal-specific tests (SP login, SP logout) are automatically skipped via it.skipIf(authMode === 'user').
  • CI isolation: CI always sets E2E_AUTH_MODE=sp with pipeline secrets. Local .env files are gitignored and never affect CI.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.