Summary
Add setup_scenarios aliases in test/e2e-scenario/nemoclaw_scenarios/scenarios.yaml for the 13 layered test_plans that currently have no user-friendly scenario ID. This unblocks deletion of the legacy scenarios/run.ts TS framework and removes a hidden coupling where the only thing that knew these IDs was the deleted listScenarios() registry.
This is Phase 4 (Onboarding Matrix Expansion) follow-up work for epic #3588. The onboarding profiles and test_plans already exist; only the user-friendly aliases are missing.
Background
While cleaning up the obsolete test/e2e-scenario/scenarios/run.ts runner (replaced by test/e2e-scenario/runtime/run-scenario.sh), we found that the OLD TS registry (scenarios/scenarios/baseline.ts) defined 23 user-friendly scenario IDs, but the new nemoclaw_scenarios/scenarios.yaml only declares 10 under setup_scenarios. The remaining 13 exist only as test_plans with internal __-form IDs.
Several places still reference the missing user-friendly IDs and currently rely on the OLD registry to resolve them:
.github/workflows/e2e-scenarios.yaml — ROUTES table maps these IDs to runner types.test/e2e-scenario-advisor.test.ts — advisor test asserts ubuntu-repo-cloud-openclaw-telegram resolves.tools/e2e-advisor/scenarios.mts — enumerates scenarios for the advisor.
Missing user-friendly IDs (13)
Each maps 1:1 to an existing test_plans entry via alias_for_plan:
| Missing setup_scenarios id |
Existing test_plans id |
ubuntu-repo-cloud-openclaw-brave |
ubuntu-repo-docker__cloud-nvidia-openclaw-brave |
ubuntu-repo-cloud-openclaw-telegram |
ubuntu-repo-docker__cloud-nvidia-openclaw-telegram |
ubuntu-repo-cloud-openclaw-discord |
ubuntu-repo-docker__cloud-nvidia-openclaw-discord |
ubuntu-repo-cloud-openclaw-slack |
ubuntu-repo-docker__cloud-nvidia-openclaw-slack |
ubuntu-repo-cloud-hermes-discord |
ubuntu-repo-docker__cloud-nvidia-hermes-discord |
ubuntu-repo-cloud-hermes-slack |
ubuntu-repo-docker__cloud-nvidia-hermes-slack |
ubuntu-repo-cloud-openclaw-resume |
ubuntu-repo-docker__cloud-nvidia-openclaw-resume-after-interrupt |
ubuntu-repo-cloud-openclaw-repair |
ubuntu-repo-docker__cloud-nvidia-openclaw-repair-existing-config |
ubuntu-repo-cloud-openclaw-double-same-provider |
ubuntu-repo-docker__cloud-nvidia-openclaw-double-same-provider |
ubuntu-repo-cloud-openclaw-double-provider-switch |
ubuntu-repo-docker__cloud-nvidia-openclaw-double-provider-switch |
ubuntu-repo-cloud-openclaw-token-rotation |
ubuntu-repo-docker__cloud-nvidia-openclaw-token-rotation |
ubuntu-repo-openai-compatible-openclaw |
ubuntu-repo-docker__openai-compatible-openclaw |
(That's 12 in the table — the 13th depends on whether ubuntu-repo-cloud-openclaw and the negative-path IDs are reconciled separately. Confirm during implementation.)
Acceptance criteria
Context — companion cleanup
This issue is the prerequisite for landing the cleanup PR that deletes the obsolete run.ts path:
- Deletes
test/e2e-scenario/scenarios/ (entire OLD TS framework, 33 files)
- Deletes
test/e2e-scenario/manifests/ (only consumed by OLD framework)
- Deletes
test/e2e-scenario/onboarding_assertions/ (dead in NEW path)
- Deletes 6 framework-tests that only test the OLD framework
- Switches
.github/workflows/e2e-scenarios.yaml from npx tsx ...run.ts to bash ...run-scenario.sh
That cleanup PR can land now (with the advisor test deferred). Once this issue lands, the advisor test will pass against the YAML resolver and the migration is fully complete.
Parent
Epic #3588 (Phase 4: Onboarding Matrix Expansion).
Summary
Add
setup_scenariosaliases intest/e2e-scenario/nemoclaw_scenarios/scenarios.yamlfor the 13 layeredtest_plansthat currently have no user-friendly scenario ID. This unblocks deletion of the legacyscenarios/run.tsTS framework and removes a hidden coupling where the only thing that knew these IDs was the deletedlistScenarios()registry.This is Phase 4 (Onboarding Matrix Expansion) follow-up work for epic #3588. The onboarding profiles and
test_plansalready exist; only the user-friendly aliases are missing.Background
While cleaning up the obsolete
test/e2e-scenario/scenarios/run.tsrunner (replaced bytest/e2e-scenario/runtime/run-scenario.sh), we found that the OLD TS registry (scenarios/scenarios/baseline.ts) defined 23 user-friendly scenario IDs, but the newnemoclaw_scenarios/scenarios.yamlonly declares 10 undersetup_scenarios. The remaining 13 exist only astest_planswith internal__-form IDs.Several places still reference the missing user-friendly IDs and currently rely on the OLD registry to resolve them:
.github/workflows/e2e-scenarios.yaml—ROUTEStable maps these IDs to runner types.test/e2e-scenario-advisor.test.ts— advisor test assertsubuntu-repo-cloud-openclaw-telegramresolves.tools/e2e-advisor/scenarios.mts— enumerates scenarios for the advisor.Missing user-friendly IDs (13)
Each maps 1:1 to an existing
test_plansentry viaalias_for_plan:ubuntu-repo-cloud-openclaw-braveubuntu-repo-docker__cloud-nvidia-openclaw-braveubuntu-repo-cloud-openclaw-telegramubuntu-repo-docker__cloud-nvidia-openclaw-telegramubuntu-repo-cloud-openclaw-discordubuntu-repo-docker__cloud-nvidia-openclaw-discordubuntu-repo-cloud-openclaw-slackubuntu-repo-docker__cloud-nvidia-openclaw-slackubuntu-repo-cloud-hermes-discordubuntu-repo-docker__cloud-nvidia-hermes-discordubuntu-repo-cloud-hermes-slackubuntu-repo-docker__cloud-nvidia-hermes-slackubuntu-repo-cloud-openclaw-resumeubuntu-repo-docker__cloud-nvidia-openclaw-resume-after-interruptubuntu-repo-cloud-openclaw-repairubuntu-repo-docker__cloud-nvidia-openclaw-repair-existing-configubuntu-repo-cloud-openclaw-double-same-providerubuntu-repo-docker__cloud-nvidia-openclaw-double-same-providerubuntu-repo-cloud-openclaw-double-provider-switchubuntu-repo-docker__cloud-nvidia-openclaw-double-provider-switchubuntu-repo-cloud-openclaw-token-rotationubuntu-repo-docker__cloud-nvidia-openclaw-token-rotationubuntu-repo-openai-compatible-openclawubuntu-repo-docker__openai-compatible-openclaw(That's 12 in the table — the 13th depends on whether
ubuntu-repo-cloud-openclawand the negative-path IDs are reconciled separately. Confirm during implementation.)Acceptance criteria
setup_scenarios:entry withalias_for_plan: <plan-id>.bash test/e2e-scenario/runtime/run-scenario.sh <each-id> --plan-onlysucceeds for all 13 IDs..github/workflows/e2e-scenarios.yamlROUTEStable is unchanged (every ID still maps).tools/e2e-advisor/scenarios.mtsenumerates these IDs from YAML (no TS registry dependency).test/e2e-scenario-advisor.test.tspasses against the YAML resolver (currently fails onubuntu-repo-cloud-openclaw-telegram).test/e2e-scenario/framework-tests/continue to pass.Context — companion cleanup
This issue is the prerequisite for landing the cleanup PR that deletes the obsolete
run.tspath:test/e2e-scenario/scenarios/(entire OLD TS framework, 33 files)test/e2e-scenario/manifests/(only consumed by OLD framework)test/e2e-scenario/onboarding_assertions/(dead in NEW path).github/workflows/e2e-scenarios.yamlfromnpx tsx ...run.tstobash ...run-scenario.shThat cleanup PR can land now (with the advisor test deferred). Once this issue lands, the advisor test will pass against the YAML resolver and the migration is fully complete.
Parent
Epic #3588 (Phase 4: Onboarding Matrix Expansion).