It shows up in the product
One task exists in PM and one Workflow Case exists for that run. You are not relying on a shell success line alone.
First proven workflow and public proof pack
Start here if you want the shortest honest OpenVibeCoding story. What can I run, inspect, and trust today before I dive into the deeper protocol and builder surfaces?
The answer is still: start with news_digest. It is the
only official release-proven public baseline. topic_brief
now has a tracked search-backed public proof bundle, and
page_brief now has a tracked
browser-backed public proof bundle, but it is still not the official
first public baseline for either path.
The 10-minute proof-first loop
- Start one `news_digest` run from the PM surface.
- Watch it show up in Command Tower and confirm the Workflow Case exists.
- Open Proof & Replay and inspect the result before trusting it.
- Reuse the Workflow Case recap as the share-ready proof asset.
If you arrived here from the agent starter kits, this page is the post-install proof target. In plain English: do not assume the starter file already proved a real integration. Come back here and make sure the same proof pack still holds.
What success should look like
You can inspect what happened
You can open proof, replay, and recap artifacts without guessing where they went or whether the run is real.
You know what is proven
You can clearly explain why news_digest is proven today and why the other two paths are not yet equivalent.
Proof status by public case
| Case | Current proof state | What to inspect today |
|---|---|---|
| news_digest | Official release-proven first public baseline | Healthy proof summary plus a tracked single-run benchmark baseline remain available inside the repo-tracked public proof bundle. |
| topic_brief | Tracked search-backed public proof bundle; not the official first public baseline | Inspect the dedicated healthy proof summary, bundle baseline summary, and share-ready Workflow Case recap now tracked under the repo-side public proof bundle. |
| page_brief | Tracked browser-backed public proof bundle; not the official first public baseline | Inspect the dedicated healthy proof summary, bundle baseline summary, and share-ready Workflow Case recap now tracked under the repo-side public proof bundle. |
The proof pack for news_digest
This is the smallest reusable trust bundle on the public surface today:
one healthy proof summary, one single-run benchmark, one Workflow Case
recap, and one storefront proof status snapshot. This page is the
human-readable proof story for that bundle. Repo-side receipts and
machine metadata still exist for maintainers and gates, but public
readers should not need to open raw ledgers or machine contracts just
to understand what is proven.
| Artifact | Why it matters | Open now |
|---|---|---|
| Healthy proof summary | Shows the tracked healthy result path for the official public baseline. | Summarized here; deeper audit metadata stays repo-side for maintainers and gates. |
| Single-run benchmark summary | Shows the current public performance baseline without pretending it is a broad benchmark campaign. | Summarized here; deeper audit metadata stays repo-side for maintainers and gates. |
| Workflow Case recap | Shows the strongest March 2026 share-ready artifact for why Workflow Cases matter outside the operator UI. | Summarized here; deeper audit metadata stays repo-side for maintainers and gates. |
| Proof status snapshot | Shows which storefront captures are healthy, degraded, missing, or still only explanation assets. | Summarized on this page instead of routed through raw status ledgers. |
Repo-side proof manifests, indexes, and capture contracts still exist for maintainers and gates, but they are not required reading for a public evaluator who just needs the shortest honest proof story.
Why this helps discovery
These use cases are not just examples for people who already know the product. They are also the easiest way to explain what OpenVibeCoding does with AI workflow cases, read-only proof surfaces, and share-ready recap assets without jumping straight into the full operator UI.
Why share-ready Workflow Cases matter
The point is not just to export JSON. The point is to make a case easier to review, easier to circulate, and easier to compare without sending everyone back through the full operator UI.
Proof you can rely on today
- The official public baseline still belongs to news_digest.
- The current benchmark story is a tracked single-run baseline, not a broad release average.
- The public recap story now includes the official news_digest recap plus tracked topic_brief and page_brief bundle assets.
- Global proof-pack index across the official baseline, tracked bundle, and showcase bundles still exists as repo-side metadata for maintainers and audits, not as a front-door jump target.
- Repo-side proof manifests, indexes, and capture contracts still exist for maintainers and gates, but they are not required reading for a public evaluator.
- This page is the public proof story; raw manifests and ledger-style contracts stay behind the scenes.
Current gaps that still matter
- topic_brief now has a tracked search-backed public proof bundle, but it is still not the official first public baseline.
- page_brief now has a tracked browser-backed public proof bundle, but it is still not the official first public baseline.
- The current benchmark story is a tracked single-run baseline, not a broad release average.
- The current recap story now has one tracked news_digest Workflow Case asset, one tracked search-backed topic_brief recap asset, one tracked browser-backed page_brief recap asset, and one remaining broader benchmark gap.
What we still do not claim
- We do not claim that topic_brief is the official first public baseline; it is not yet equally release-proven with news_digest.
- We do not claim that page_brief replaced news_digest as the official first public baseline; it is not yet equally release-proven with it.
- We do not claim that a storyboard or degraded local capture counts as healthy end-to-end proof.
- We do not claim that the current benchmark file is a broad multi-run release average.
- We do not claim that a public evaluator should read raw machine metadata or maintainer-grade capture contracts to understand the proof story.
Need the surrounding operator and builder story? Visit the compatibility matrix, the AI + MCP + API surface guide, the builder quickstart, or the ecosystem map.