AG
agentra-e2e-testing
Plan, write, run, or review Agentra E2E tests using Playwright/browser tooling. Use for critical chat/auth/user journeys, flaky tests, screenshots, traces, and browser-driven validation.
Install
mkdir -p .claude/skills/agentra-e2e-testing && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14816" && unzip -o skill.zip -d .claude/skills/agentra-e2e-testing && rm skill.zipInstalls to .claude/skills/agentra-e2e-testing
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Plan, write, run, or review Agentra E2E tests using Playwright/browser tooling. Use for critical chat/auth/user journeys, flaky tests, screenshots, traces, and browser-driven validation.186 chars✓ has a “when” trigger
About this skill
Agentra E2E Testing
Use this skill for browser-level validation of Agentra workflows.
Sources of truth
AGENTS.md- Existing Playwright/browser-use tooling in the Codex environment.
.claude/skills/e2e-testing/SKILL.mdremains a reference for generic E2E patterns when deeper examples are needed.
Workflow
- Identify the critical user journey and owning workspace.
- Prefer existing app scripts and test patterns before introducing new test structure.
- Use stable selectors and explicit waits based on user-visible state.
- Capture screenshots, traces, or console/network details when diagnosing.
- Quarantine or document flaky behavior instead of hiding it with broad waits.
Agentra focus areas
- Auth redirects and signed-in/signed-out states.
- Chat request lifecycle and SSE/stream interruption behavior.
- Error, loading, empty, and cancellation states.
- Runtime smoke handoff when UI symptoms point to AgentCore behavior.
Output
Report journey, command or browser target, artifacts captured, result, and follow-up risk.