/gaia-test-e2e

user-facing
Category:
Testing
Lifecycle phase:
5 -- Deployment
Arguments:
[story-key] [--adapter <name>] [--target-url <url>]

What it does

/gaia-test-e2e executes end-to-end tests via a Playwright or Cypress adapter against a deployed environment. The Test Execution Bridge runs the adapter (Phase 3A), then applies LLM judgment on the results to assess test stability, coverage adequacy, and failure root causes (Phase 3B). The final verdict is computed by the verdict resolver -- the LLM does not determine the verdict.

When to use it

  • You have deployed to staging or production and want to run the e2e test suite against the live environment.
  • You want to verify critical user paths after a deployment.

Prerequisites

  • An e2e adapter must be configured via test_execution.e2e.adapter in .gaia/config/project-config.yaml, or pass --adapter.
  • The underlying tool must be installed (Playwright or Cypress). If missing, the command reports BLOCKED with installation guidance.

How to invoke

/gaia-test-e2e E3-S7
/gaia-test-e2e --adapter cypress
/gaia-test-e2e --target-url https://staging.example.com
/gaia-test-e2e --adapter cypress --target-url https://staging.example.com

What it does step by step

  1. Resolve adapter Selects the e2e adapter: --adapter flag, then config, then default (playwright-e2e).
  2. Probe tool Checks that the underlying tool is installed. Reports BLOCKED if missing.
  3. Execute tests (Phase 3A) Runs the adapter's test suite and captures structured results.
  4. LLM judgment (Phase 3B) Evaluates test stability, coverage adequacy, and failure root causes against the e2e rubric.
  5. Verdict resolution Computes APPROVE, REQUEST_CHANGES, or BLOCKED from deterministic checks plus LLM findings.
  6. Update Review Gate If a story key was provided, maps the verdict to PASSED or FAILED in the Review Gate.

Inputs

InputSourceDescriptionExample
story-keyArgument (optional)Links the run to a story's Review Gate./gaia-test-e2e E3-S7
--adapterArgument (optional)Override the configured e2e adapter./gaia-test-e2e --adapter cypress
--target-urlArgument (optional)Override the target URL./gaia-test-e2e --target-url https://staging.example.com

Outputs

OutputLocationDescription
Analysis resultsevidence/e2e/Structured adapter output with test counts, pass/fail, and timing.
VerdictConversation + Review GateAPPROVE, REQUEST_CHANGES, or BLOCKED.

Example session

> /gaia-test-e2e E3-S7 --target-url https://staging.example.com

Adapter: playwright-e2e (available)
Running e2e suite...

12 tests passed, 1 failed, 0 skipped (18.4s)

Failed: login-redirect.spec.ts -- assertion timeout
Root cause: likely application bug (not infrastructure flake)

Verdict: REQUEST_CHANGES
Review Gate updated: FAILED

What to run next

  • Fix the failing test or underlying bug, then re-run.
  • /gaia-post-deploy -- run extended post-deployment verification.

Troubleshooting

BLOCKED (tool not installed)

Install the underlying tool (Playwright or Cypress) and re-run.

Target URL unreachable

Verify the deployment is running and the URL is correct.