Files
openclaw/qa/scenarios/runtime/docker-system-agent-first-run.yaml
T

34 lines
1.7 KiB
YAML

title: Docker packaged OpenClaw inference gate
scenario:
id: docker-system-agent-first-run
surface: containers
category: containers.container-setup
coverage:
secondary:
- containers.first-run-onboarding
- small-linux.first-run-verification
objective: Verify a fresh package-installed Docker state gates OpenClaw on a tested inference route, then supports planner-backed and typed setup operations without leaking secrets.
successCriteria:
- The bare-command routing policy selects onboarding, and the packaged OpenClaw CLI exits with onboarding guidance while inference is unavailable.
- The packaged activation module tests fake Claude before persisting its model, while workspace and Gateway setup remain untouched.
- The modern compatibility entrypoint becomes available after activation.
- A fuzzy request uses the verified planner for typed setup; subsequent one-shot commands write the default workspace/model and create the configured agent.
- Discord setup is written through a SecretRef without persisting the raw token.
- Config validation and expected OpenClaw audit entries succeed.
docsRefs:
- docs/install/docker.md
- docs/cli/setup.md
- docs/help/testing.md
codeRefs:
- scripts/e2e/system-agent-first-run-docker.sh
- test/e2e/qa-lab/runtime/system-agent-first-run-docker-client.ts
- scripts/e2e/system-agent-first-run-spec.json
execution:
kind: script
path: test/e2e/qa-lab/runtime/docker-e2e-lane.ts
summary: Runs supporting packaged-CLI evidence for the inference gate, activation module, planner, and typed operations; it does not claim an interactive onboarding or OpenClaw agent/tool/approval flow.
args:
- --lane
- system-agent-first-run