Make Claude Code work
the way you do.

# Using your team's pytest layout
# (python-testing skill loaded)

@pytest.mark.asyncio
async def test_create_user_with_valid_email(
    db_session, sample_user_data
):
    user = await create_user(db_session, **sample_user_data)
    assert user.email == sample_user_data["email"]
    assert user.id is not None

@pytest.mark.parametrize("invalid", [
    "", "no-at-sign", "@no-local-part"
])
async def test_create_user_rejects_invalid_email(
    db_session, invalid
):
    with pytest.raises(ValueError, match="invalid email"):
        await create_user(db_session, email=invalid)

# de-slopify skill loaded

Subject: See exactly where your signups drop off

The new analytics view answers one question fast: which
step of your funnel loses the most people. Open it, sort
by drop-off rate, and fix the worst step first.

Live now in Settings -> Analytics.

# ab-testing skill loaded

1. Sample size: ~1,400 / variant for 80% power to
   detect a 10% lift at p < 0.05.
2. Server-side assignment in middleware (no flicker),
   variant pinned by cookie + logged to GA4.
3. Bayesian readout: ship when P(B > A) > 95%.
4. Stops at a human checkpoint before routing traffic.

# ga4 skill loaded — event taxonomy

gtag('event', 'sign_up', {
  method: 'email',
  funnel_step: 'completed',
  plan: selectedPlan,
});
// + Consent Mode v2 and a server-side Measurement
// Protocol backup, validated in GA4 DebugView