CR R1 follow-ups on top of the /api/ops proxy fix.
Bucket (a) — must-fix:
- Dockerfile: declare ARG/ENV OPS_BASE_URL in the builder stage. next.config.ts
evaluates rewrites() at build time and throws if OPS_BASE_URL is unset, which
aborted `next build` in CI. Mirrors the existing NEXT_PUBLIC_SHELL_URL /
NEXT_PUBLIC_POCKETBASE_URL pattern.
- showcase_deploy.yml: pipe OPS_BASE_URL through to docker build for the
shell-dashboard matrix entry, defaulting to the production
showcase-ops-production.up.railway.app URL.
- next.config.ts: strip trailing slashes from OPS_BASE_URL before constructing
the rewrite destination, matching the same normalization in
src/lib/ops-api.ts:resolveBaseUrl so server-side rewrite and client-side
fetch agree on the URL shape.
- src/lib/ops-api.ts: treat empty / whitespace NEXT_PUBLIC_OPS_BASE_URL as
"no override". `??` only short-circuits on null/undefined, so an env var set
to "" silently produced baseUrl="" and URLs of the form "/probes" with no
/api/ops prefix.
- use-probes.integration.test.tsx: snapshot+restore process.env.NEXT_PUBLIC_OPS_BASE_URL
in beforeEach/afterEach so tests never leak env state. Strengthen the proxy
contract assertions: lock toHaveBeenCalledTimes(1), assert method=GET,
cache=no-store, accept JSON header, and signal is an AbortSignal. Tighten
the 404 regression to assert the canonical ensureOk message shape so a
refactor that changes the format trips the test.
Bucket (b) — applied since the diff stayed focused:
- triggerProbe: add cache:"no-store" for parity with the GET fetches.
- fetchProbeDetail / triggerProbe: throw early when id is empty so callers
get a clean error instead of a request to /probes//... .
- ensureOk: bump body-truncation cap from 200 to 500 chars and append a
`[truncated, N bytes total]` marker so operators can see they're missing
tail bytes when the server returns a long HTML/stack-trace body.
- ops-api.ts: drop dev-loop review-cycle tag prefixes (R2-C.3, R3-C, R3-D.1)
from comments. Keep the actual rationale.
- ops-api.ts header docstring: clarify that NEXT_PUBLIC_OPS_BASE_URL is read
live at runtime in this codebase (SSR + tests), not just statically inlined
into the client bundle.
Verified:
- Tests: vitest run — 26 files, 293 passed, 1 skipped (no test count change).
- Typecheck: tsc --noEmit clean.
- Lint + format: oxlint + oxfmt clean on changed files.
- Local Docker build: `docker build --build-arg OPS_BASE_URL=https://...` succeeds.
Without --build-arg the build fails with "OPS_BASE_URL must be set" as
expected, confirming the fix is load-bearing.
showcase-ops was excluded from .github/workflows/showcase_deploy.yml, so
commits touching showcase/ops/** never produced a fresh GHCR image. PR
#4293 (Status tab + /api/probes route) merged to main on 2026-04-26 but
no rebuild fired — the deployed Railway image is stale and /api/probes
404s in production.
Adding showcase-ops as a first-class matrix entry:
- dispatch_name: showcase-ops (workflow_dispatch option + filter_key)
- paths-filter: showcase/ops/**, plus shared/scripts/manifests
(showcase-ops's Dockerfile bundles all four into the runtime image
via build-stage COPY + generate-registry.ts)
- context: '.' (repo root) so the Dockerfile can COPY from
pnpm-workspace.yaml + packages/ + showcase/{ops,shared,packages,scripts}
- dockerfile: showcase/ops/Dockerfile
- image: showcase-ops -> ghcr.io/copilotkit/showcase-ops:latest
- railway_id: 3a14bfed-0537-4d71-897b-7c593dca161d
- health_path: /health (matches Dockerfile HEALTHCHECK + Hono route)
- timeout: 20 (heavier build than shells: pnpm deploy + chromium
install via playwright --with-deps)
- lfs: false (no Git LFS assets in showcase/ops)
- linux/amd64 platform inherited from existing build step (Depot)
Resulting matrix: 39 services (was 38). dispatch_name uniqueness +
JSON validity verified locally; actionlint/yamllint surface only
pre-existing findings on the workflow.
oxfmt 0.36 supports .md but not .mdx — including .mdx in the file-list
glob causes MDX-only PRs to fail with "Expected at least one target
file" because oxfmt drops every input as an unknown target and then
errors on the empty target set. Mixed PRs (.mdx + .tsx/.json/etc) pass
because the non-MDX files keep the target set non-empty, which is why
this has only surfaced now on a shell-docs-only sync PR.
Removing .mdx from the glob lets MDX-only PRs hit the existing count=0
skip path and pass cleanly. Add .mdx back when oxfmt ships MDX support.
The test at integration-smoke.spec.ts:21 imports registry.json, which
is gitignored (generated at build time). After PR #4236 removed it
from tracking, every CI run fails with "Cannot find module
'../../shell/src/data/registry.json'" — producing the repeating
"Starter Deployed Smoke Test Failed — 0 failure(s) — job-level error"
Slack alerts in #oss-alerts.
Adds a generate-registry step before the Playwright test run.
Three fixes batched to minimize PR churn:
1. Dockerfile: add ARG NEXT_PUBLIC_POCKETBASE_URL so the PB URL gets
baked into the Next.js bundle at build time. Without this, pb.ts
resolves to the sentinel URL and the dashboard shows "unavailable"
on every tab. Pre-existing bug exposed by fresh deploys.
2. showcase_deploy.yml: pass NEXT_PUBLIC_POCKETBASE_URL and
NEXT_PUBLIC_SHELL_URL as build args for the shell-dashboard service
in the CI matrix. Neither was ever passed before.
3. cell-matrix.tsx + parity-matrix.tsx: flatten nested table pattern
that caused column misalignment. Category rows used colSpan with
an inner <table> whose columns floated independently of the header.
Replaced with useCollapsible hook + flat sibling <tr> rows.
Also regenerates package-lock.json for the plugin-react downgrade
from PR #4241 (npm ci was failing in Docker).
Local Docker build verified with --build-arg for both NEXT_PUBLIC vars.
- deploy workflow: add shared/scripts/manifest paths to shell-dashboard
and shell-docs filters (previously triggered implicitly by committed
JSON diffs in those directories)
- capture-previews: add generate-registry step before capture; use
git add -f for the gitignored registry.json
- e2e smoke test: document generator dependency in import comment
Two bugs in the earlier version, both surfaced once main advanced and
was merged into the branch:
- `fetch-depth: ${{ ... && 0 || 1 }}` evaluated to `1` on PRs because
the short-circuit treats `0` as falsy, so the base SHA was missing
locally and `git diff` exited 128.
- Diffing against the PR's stored `base.sha` includes every file main
touched since the PR opened once main is merged into the branch, which
defeats the whole point of this change. Diff against the current tip
of the base branch instead.
The static / quality "format" job ran oxfmt across the whole tree on
every PR. When main advanced with stale files, unrelated PRs picked up
those re-formats and ended up with a noisy `style: auto-fix formatting`
commit.
Now the job collects `git diff --name-only base...HEAD` for oxfmt-
supported extensions and passes only those paths to oxfmt. Push events
on main continue to check the whole repo so the baseline is still
enforced.
aimock is no longer a Docker-built showcase service. Railway
pulls the pre-built upstream image directly from GHCR. Remove:
- aimock from workflow_dispatch service options
- aimock paths-filter (showcase/aimock/**)
- aimock entry from ALL_SERVICES matrix
The capture job can take ~30 minutes, during which other commits
routinely land on main. Without a rebase-and-retry loop, the
devops-bot push loses the race and fails with "fetch first".
Observed in runs 24799601181 and 24809331709 on 2026-04-22,
both failing at the same step with identical remote-rejected-push
output. Bounded to 5 attempts so a persistent failure still
surfaces rather than looping forever.
## Summary
- Change `shell-dojo` matrix entry: `context: "."` + `dockerfile:
"showcase/shell-dojo/Dockerfile"`. Matches shell-dashboard pattern.
## Why
Build fails with `failed to walk /tmp/buildkit-mount.../showcase: no
such file or directory`. `showcase/shell-dojo/Dockerfile` has `COPY
showcase/scripts/...`, `COPY showcase/shared/...`, `COPY
showcase/packages/...` — expects repo root as build context. Matrix had
`context: showcase/shell-dojo` which scopes the context to just
shell-dojo's own directory, making those COPYs fail.
## Test plan
- [ ] Next deploy of shell-dojo succeeds
## Summary
- Replace `jq -n --rawfile text … > \$(mktemp)` + `payload-file-path`
with inline `payload:` + `toJSON(format(...))` on the scheduled
Slack-alert step in `test_smoke-starter.yml`.
- Pattern already in use in `test_smoke-starter-deployed.yml` (PR #4068)
and `showcase_validate.yml` — this brings the last holdout in line.
## Why
Every scheduled run of `test / smoke / starter` that legitimately fails
a matrix leg (upstream outage, floating-dep breakage, etc.) has been
crashing the downstream alert step with:
```
##[error]Invalid input! Failed to parse contents of the provided payload file
```
Two compounding causes:
1. `slackapi/slack-github-action@v2.1.0` rejects payload files whose
name does not end in `.json`/`.yaml`/`.yml`. `mktemp` produces
extensionless files.
2. `jq -n --rawfile text "$SLACK_MSG"` under `set -e` can abort on
edge-case summary input (missing tmpfile, non-UTF-8 bytes surviving the
sanitizer), leaving an empty or missing payload.
Either way, the notifier failed before firing, so genuine outages went
unreported in `#oss-alerts`. The user reported this as "Showcase: Drift
Detection" — that specific workflow was already retired to showcase-ops
on 2026-04-22 (commit `89eb0734b`), but the same
jq-rawfile-then-payload-file-path pattern still lived in
`test_smoke-starter.yml`.
Inline `payload:` with `toJSON(format(...))` sidesteps both failure
modes: quotes, backslashes, and newlines in the sanitized summary are
safely JSON-encoded at template-eval time, and there is no intermediate
file to mishandle.
Audit of remaining file-based payloads left intentionally unchanged:
- `test_smoke-starter-deployed.yml` recovery step writes to
`/tmp/starter-smoke-recovery.json` (valid extension, simple `jq --arg`
scalar) — safe.
- `showcase_docs-sync.yml` writes to `slack-payloads/*.json` (valid
extensions, `jq --arg` scalars) — safe.
## Test plan
- [ ] `actionlint .github/workflows/test_smoke-starter.yml` reports no
new findings (confirmed locally — one pre-existing line-64 warning
unrelated to this diff)
- [ ] Next scheduled run on `main` that has a genuinely failing matrix
leg posts a well-formed red alert to `#oss-alerts` instead of `Invalid
input! Failed to parse contents of the provided payload file`
- [ ] A scheduled run where all matrix legs pass remains silent (no
behavior change on the green path)
The failure-alert Slack step crashed every scheduled run with
'Invalid input! Failed to parse contents of the provided payload file'.
Two compounding causes: slackapi/slack-github-action@v2.1.0 rejects
payload files without a .json/.yaml/.yml extension (mktemp produces
extensionless files), and 'jq -n --rawfile' under 'set -e' can abort
on edge-case summary input, leaving an empty/missing payload.
Switch to the inline 'payload:' + toJSON(format(...)) pattern already
used in test_smoke-starter-deployed.yml (PR #4068) and
showcase_validate.yml. Summary text is sanitized into $GITHUB_ENV via
heredoc (handles embedded =, quotes, newlines), then JSON-encoded by
toJSON at template-eval time — no intermediate file, no jq crash path,
genuine outages surface in #oss-alerts instead of being silently
suppressed by a broken notifier.
test_smoke-starter-deployed.yml recovery path (/tmp/...-recovery.json,
jq --arg) and showcase_docs-sync.yml (static slack-payloads/*.json)
use safe patterns already — left unchanged.
GitHub's "Create a merge commit" merge option builds the commit message
from the PR body. When the body contains markdown lists or blank lines
(e.g. PR #4113 → merge 5ed233f01), commitlint parses subsequent
paragraphs as additional commit subjects and fails with subject-empty /
type-empty.
Two-layer fix:
- commitlint.config.js: ignore standard "Merge " prefixed messages.
- static_quality.yml: guard the push-path `--last` step with a
parent-count check so true merge commits (which keep the PR-title
header and thus don't match the "Merge " prefix) are skipped
before commitlint runs at all.
The workflow runs exclusively on main-branch pushes, completed
"Showcase: Build & Deploy" runs, and manual dispatch — all production
events where silent failures (e.g. the GH013 PROTECT_OUR_MAIN
regression that motivated PR #4159) must surface in #oss-alerts
rather than getting buried in the Actions tab.
Mirrors the failure-alert pattern from showcase_validate.yml:
- Hoist SLACK_WEBHOOK_OSS_ALERTS into a job-level env var so
step-level `if:` expressions can reference it (secrets.* is
not a valid named-value inside `if:`).
- Best-effort "Extract failure details" step pulls the failed
step name and first meaningful error line from the jobs API +
`gh run view --log-failed`, truncated to 300 chars.
- `slackapi/slack-github-action@v2.1.0` with toJSON(format(...))
wrapping to safely JSON-encode any dynamic values.
- Fallback `::warning::` log when the webhook secret is unset so
the gap is still visible in the workflow output.
Gated on `failure() && env.SLACK_WEBHOOK != ''`. No `github.event_name`
filter needed — this workflow has no pull_request trigger, so every
failure is an actionable production event.
Adds a new step that runs validate-fixture-tool-surface.ts on every PR
and push to main. Sits alongside the existing validate-parity /
validate-workflow-starters / validate-pins steps and follows the same
pnpm-exec-tsx pattern.
Without this, the drift validator only runs locally or via the
vitest suite (which only catches bugs in the validator itself, not
drift in the real fixture/demo state). The CLI invocation against the
committed tree is what would have caught the 2026-04-22 regression
before it reached prod.
## Summary
- **CI workflow consolidation**: Merged commitlint into static_quality,
added concurrency groups to test workflows, added path filters to
publish-commit, fixed static_danger matrix bug and python-sdk
concurrency
- **Showcase deploy refactor**: Replaced 23 duplicate jobs with a
dynamic matrix strategy (~1000 lines removed)
- **Preview capture → MP4**: Switched from GIF to optimized MP4 (H.264,
CRF 28, faststart), uploads to GitHub Release `showcase-previews`
instead of committing blobs to the repo. DemoCard uses `<video autoplay
muted loop>` for hover previews.
## Test plan
- [x] Local MP4 capture test passed (langgraph-python, 14KB, 8s, H.264
400x300 @ 10fps)
- [ ] CI workflows pass on this branch
- [ ] Capture workflow produces and uploads MP4 to release
- [ ] DemoCard hover shows video preview when URL is populated
🤖 Generated with [Claude Code](https://claude.com/claude-code)
## CI hygiene fixes
Four direct-fix items surfaced during the 2026-04-16 QA/E2E blitz,
bundled as one PR with one commit per fix:
1. **`test_unit.yml` paths-ignore** — add `showcase/**` +
`sdk-python/**` (prevents spurious TS unit matrix runs on showcase-only
or sdk-python-only PRs). `sdk-python-old/**` was mentioned in the plan
but doesn't exist on main; only the two existing dirs are added.
2. **`test_doc-examples.yml`** — scope PR trigger to `branches: [main]`
(matches convention of other workflows).
3. **`e2e_dojo.yml`** — symmetric `.changeset` path filter on push+PR
(was asymmetric — already in the `dorny/paths-filter` step's `ts:` list,
so this aligns the trigger).
4. **`starter-smoke.yml`** — rename internal job id (`starter-smoke` →
`smoke-starter`) and artifact name pattern (cosmetic; matches the new
`test_<layer>-<target>` / `smoke-<layer>` naming convention; no external
consumers).
### Fix skipped
**`showcase_smoke-monitor.yml` slug normalization** (item #10 in the
Notion page) — inspecting the file, it already uses the
`showcase/packages/` slug convention (`ms-agent-python`,
`ms-agent-dotnet`, `strands`). The apparent inconsistency is actually in
`starter-smoke.yml`, whose matrix keys must stay as
`ms-agent-framework-*` / `strands-python` because they are literal
directory names in `examples/integrations/` (used as `working-directory:
examples/integrations/${{ matrix.starter }}`). So there is no actionable
change here — the two files use different slug conventions *by
necessity*, because they target different directories (deployed
`showcase/packages/*` vs. local `examples/integrations/*`). Flagging for
the author of the blitz notes in case the actual concern was something
else.
Refs: [Bugs Found During
Blitz](https://www.notion.so/3443aa381852812fb595c5118dd68818) items #3,
#8, #9, #12.
## Summary
- Rewrite starter Dockerfile templates (`typescript`, `python`, `java`,
`dotnet`) as true multi-stage: builder has dev tools, runner is minimal.
No `pip`/`npm`/`pnpm`/`tsx`/`*-cli dev` in runtime stages.
- Encode prod-mode behaviors for `claude-sdk-typescript`, `mastra`,
`langgraph-typescript` directly in the TS template + generator
conditionals (folds in the pattern from #4132 for langgraph-ts; adds
parallel migration for claude-sdk-ts + mastra).
- Regenerate all 17 starter Dockerfiles.
- Rewrite all 17 package Dockerfiles to match the template shape
(multi-stage, venv builder for Python, 3-stage Python runtime, `USER
app`, etc.).
- Cold-start instrumentation on `claude-sdk-typescript` package
(absorbed from closed PR #4133).
- Pin `platforms: linux/amd64` in the showcase deploy workflow
(`docker/build-push-action` step) to match what Railway + GHCR require.
- Template `README.md` documenting multi-stage conventions, size target,
platform mandate, and the `dockerfile_hygiene` probe.
- `generate-starters.ts` strips the `langgraph-typescript` starter's
dead `server.mjs` + its `start` script (both reference deps only needed
in the package's prod-mode path).
## Why
Railway cold starts exceed the 180s watchdog grace when the runtime
stage boots via `langgraph-cli dev` / `npx mastra dev` / `npx tsx` — the
dev-mode work (TS compile, Rollup bundle, native libsql init) blocks the
health probe. #4132 proved the fix is to move this to image build time
for langgraph-ts. This PR generalizes that shape to all TS starters that
need it and reshapes every Dockerfile in the tree so runners no longer
carry dev toolchains.
Secondary outcomes:
- Smaller runtime images (pip/npm toolchain moved to builder only).
- Consistent `--platform linux/amd64` pinning end-to-end.
- No `pnpm install --frozen-lockfile` in the TS starter template
(starters are no-lockfile by design).
- `validate-pins` ratchet re-baselined to match the regenerated spec
strings.
## Test plan
- [ ] `docker build --platform linux/amd64` green for every starter (17)
and package (17) locally
- [ ] `pnpm -C showcase/scripts` vitest: 1085/1085 pass
- [ ] `tsc --noEmit -p showcase/scripts/tsconfig.json` clean
- [ ] `oxfmt --check` + `oxlint` clean on touched TS files
- [ ] validate-pins: count 110, hash matches updated baseline
- [ ] CI Showcase Validate green on PR HEAD
- [ ] Post-merge: showcase deploy for mastra + claude-sdk-typescript
goes green (currently flapping)
## Why
Adds an interim GitHub Actions cron that curls `/api/health` on each of
the 17 showcase starter services every 5 minutes. The goal is to keep
the agent's in-memory state warm — JIT tiers, module cache, connection
pools. Railway Pro tier doesn't sleep containers, but warm-state still
decays over idle periods.
This is a stopgap until the `showcase-ops` service (see Notion proposal
§2a) lands. Its smoke probe runs at the same cadence and will subsume
this workflow entirely.
## What
New file: `.github/workflows/showcase_keep-alive.yml`
- Schedule: `*/5 * * * *` (every 5 minutes) + manual `workflow_dispatch`
- Matrix over all 17 starter slugs (`ag2`, `agno`, `claude-sdk-python`,
`claude-sdk-typescript`, `crewai-crews`, `google-adk`,
`langgraph-fastapi`, `langgraph-python`, `langgraph-typescript`,
`langroid`, `llamaindex`, `mastra`, `ms-agent-dotnet`,
`ms-agent-python`, `pydantic-ai`, `spring-ai`, `strands`)
- Each cell: `curl -fsS --max-time 15
https://showcase-<slug>-production.up.railway.app/api/health`
- `fail-fast: false` + `continue-on-error: true` so one starter being
down doesn't cancel the others
- No Slack notification — `showcase_smoke-monitor.yml` already covers
alerting. This workflow is strictly keep-alive, not observability.
- Header comment marks it INTERIM and references the showcase-ops
cutover
## Interim marker
Remove this workflow once `showcase-ops` ships its smoke probe at the
same 5-min cadence (tracked in the Notion showcase-ops proposal).
## Test plan
- [ ] Verify workflow parses in the Actions UI after merge
- [ ] Trigger `workflow_dispatch` once manually; confirm all 17 cells
run and pass for healthy starters
- [ ] Confirm no Slack noise generated by this workflow
- [ ] Remove when showcase-ops cutover lands
Add platforms: linux/amd64 to the depot/build-push-action invocation in
showcase_deploy.yml. Railway and GHCR serve x86 hosts, so an arm64-only
image crashes on pull. Matches the platform requirement enforced for
local docker build invocations (documented in
showcase/starters/template/README.md).
Verify Marketplace credential expects an active Azure session via azure/login, but login was gated only on published==false. When dry_run=true and version is already published, the verify step ran without a session and failed. Widening the guard to include dry_run.
The dry-run summary referenced scripts/release/vscode-extension-release.sh,
which was removed in 17c4daee3 (drop release helper script). Point
maintainers at the new CHANGELOG-edit flow instead.
Adds .github/workflows/vscode-extension-changelog-sync.yml: on PRs that
touch packages/vscode-extension/CHANGELOG.md, read the top '## X.Y.Z'
entry and bump packages/vscode-extension/package.json to match via an
auto-committed 'chore: release vX.Y.Z' on the PR branch.
Collapses the release maintainer flow to a single CHANGELOG edit.
Forks are skipped (GITHUB_TOKEN can't push to fork branches). A
guardrail rejects CHANGELOG entries older than the current package.json
version (catches history edits vs. prepends).
Lets us validate the Marketplace-OIDC and Open VSX auth chains end-to-end
without actually publishing. workflow_dispatch now accepts a dry_run input
(default true); when set, the job runs checkout, Azure login,
vsce verify-pat, and a new Open VSX /api/user token introspection, then
stops before any publish, tag, release, or Slack success notify.
Push trigger is unchanged — dry_run is only meaningful for manual runs.
- Use slackapi/slack-github-action@v2.1.0 matching 8+ existing workflows
- Switch from SLACK_WEBHOOK to SLACK_WEBHOOK_OSS_ALERTS (the secret
actually wired up in the repo, routed to #oss-alerts)
- Guard with env.SLACK_WEBHOOK_OSS_ALERTS != '' for repo-fork safety
- Add failure-path notification with failed job + per-registry outcomes
+ run URL (follows memory rule: red alerts carry actionable detail)
- Expose publisher as a step output to construct Open VSX URL cleanly
(https://open-vsx.org/extension/<publisher>/<name>) instead of inline
tr substitution
Replace tag-triggered publish with push-to-main + version-on-Marketplace
self-gate, matching CopilotKit/aimock. CI reads version from package.json,
queries vsce show for that version, and no-ops if already published. On a
new version it builds once, dual-publishes (Marketplace + Open VSX with
the existing retry wrappers and idempotent 'already exists' handling),
tags vscode-extension-vX.Y.Z, cuts a GitHub Release from the CHANGELOG
section, and posts to SLACK_WEBHOOK if configured. Path-scoped to
packages/vscode-extension/** so unrelated pushes don't trigger the job.
Wrap both registry publish steps in a bash retry helper that retries up
to 5 times with staggered backoff (10s/20s/40s/60s/90s) on transient
conditions (5xx, timeouts, connection resets, DNS). Auth and validation
errors still fail fast with no retry.
Critically, 'version already exists' is treated as idempotent success:
if attempt N-1 landed on the registry but its response was lost to a
502 after commit, attempt N sees the already-published version and
short-circuits rather than failing the job.
Motivated by Open VSX /publish returning intermittent 502 Bad Gateway
errors from Eclipse Foundation infra. Each attempt is wrapped in
::group:: markers so per-attempt logs are collapsible in the Actions
UI. Reconciliation step updated to reflect retry semantics and to
tell the operator to rerun the job (not bump the version) on exhausted
retries. RELEASING.md gets a 'Transient registry failures' section
documenting the behavior for both CI and manual publish paths.
Trigger: push of vscode-extension-v* tags.
Builds the extension once, packages a single VSIX, uploads it as a
workflow artifact, then publishes the same VSIX to the VS Code
Marketplace and Open VSX. Both registry publishes use continue-on-error
with a final reconciliation step so a partial failure is visible
instead of silently succeeding on one registry only.
Secrets VSCE_PAT and OVSX_PAT are read from the 'production'
environment so rotation and approval flow are scoped correctly.
## Summary
Three connected features land together so the CopilotKit VS Code
extension becomes a coherent debugger/preview surface:
1. **Hook Explorer** — every V1 + V2 render hook can be discovered and
previewed live with auto-generated controls, an inline `▶️ Preview
Component` CodeLens, and a sidebar that lists every captured site.
2. **AG-UI Event Inspector** — live SSE debug stream of all AG-UI
events, filterable and color-coded, in a sidebar view + editor panel.
3. **A2UI Catalog sidebar → webview** — the last native TreeView gets
replaced with a Tailwind-styled webview that matches the other two, now
with a proper **Go to source** action on components and fixtures.
## Hook Explorer
### Discovery + preview
- oxc-based scanner walks the workspace and finds every call-site of any
hook in the registry (17 across V1 + V2, render + data).
- Preview panel bundles the user's source via Rolldown (IIFE format,
React externalized, CSS collected per `@copilotkit/a2ui-renderer`
pattern), executes it in the webview with a capture-only **stub** for
`@copilotkit/react-core` (+ `/v2`), and mounts the user's component just
long enough to record each hook's config.
- Auto-generated form on the left/top drives the `render` prop's
args/parameters/state/event live. V1 parameter arrays and V2 Zod /
Standard Schema all map through a unified `FormSchema` derived at
runtime from the captured config.
- `useCopilotAction`, `useCopilotAuthenticatedAction_c`,
`useCoAgentStateRender`, `useLangGraphInterrupt`, `useRenderTool`,
`useRenderToolCall`, `useDefaultRenderTool`, `useLazyToolRenderer`,
`useRenderCustomMessages`, `useRenderActivityMessage`,
`useHumanInTheLoop`, `useInterrupt`, `useFrontendTool`, `useComponent`,
`useDefaultTool` all previewable.
- Inline `▶️ Preview Component` CodeLens above every render-hook call
site, backed by the same `copilotkit.hooks.preview` command as the
sidebar.
- Imported render components work: rolldown walks transitive imports
from the hook's `render` prop through any number of sibling files.
- Cross-file hook switches are robust: controls are reset on load,
Harness only mounts once the real HostRoot is ready, a top-level error
boundary auto-recovers when you pick a different hook.
### Why the stub approach
Bundling the real `@copilotkit/react-core` through rolldown's IIFE
output hit a `__commonJSMin` TDZ chain (`require_clipboard`,
`require_graphql`, `require_context_helpers`, …) because the
chat/runtime-client/markdown graph has circular imports. Externalizing
react-core + routing to a Proxy-backed stub that captures hook configs
avoids the whole CJS wrapping problem, shrinks the preview bundle from
~24 MB to ~1.3 KB, and keeps the preview runtime path completely
runnable without a live CopilotKit backend. Tradeoff documented in
`copilotkit-stubs.ts`.
### Weather-themed fixtures
14+ fixtures under `packages/vscode-extension/test-workspace/hooks`,
each a distinct visual scenario (forecast card, severity-palette alerts
with imported CSS, forecast strip, live radar grid, conic-gradient
precipitation gauge, air-quality badge with imported render, pollen
report with a 2-hop import graph, HITL evacuation confirm,
sunrise/sunset gradient, etc.). Used both as regression fixtures and as
the demo surface for video.
### Styling
- Tailwind-via-CDN + VS Code CSS variables for theme-aware chrome.
- User-provided CSS imports collected by rolldown and injected as a
`<style>` tag per load.
- Controls + form fields converted to Tailwind; textarea matches input
styling.
- Framed "Rendered output" card so the render prop is visually
unmistakable.
## AG-UI Event Inspector
### Runtime (`@copilotkit/runtime` + `@copilotkit/shared`)
- `DebugEventBus` — in-memory pub/sub on `BaseCopilotRuntime`, only
instantiated when `NODE_ENV != production`.
- Event tap in `createSseEventResponse` broadcasts every AG-UI event
with metadata (agentId, threadId, runId, timestamp).
- `GET /debug-events` SSE endpoint — returns 404 in production, streams
`DebugEventEnvelope` JSON to connected clients, initial `: connected`
comment flushes headers immediately.
### VSCode Extension
- `DebugStream` — Node SSE client with auto-reconnect, exponential
backoff, URL validation, error surfacing.
- `InspectorPanel` (editor panel, command `CopilotKit: Open AG-UI
Inspector`) and `InspectorViewProvider` (sidebar view) both use a shared
`DebugStream` instance — events persist when switching tabs.
- Inspector React app: `ConnectionBar`, `FilterBar`, `EventList`,
`EventDetail`.
- Color scheme: purple (lifecycle), red (errors), blue (text), orange
(tools), green (reasoning), teal (state), yellow (activity), gray
(unknown).
## A2UI Catalog → webview
- Replaces `ComponentPreviewProvider` (native TreeDataProvider) with
`CatalogListViewProvider` (WebviewViewProvider), matching the Hooks and
Inspector sidebars.
- New React webview with refresh header, component rows (name + relative
path + `auto` badge when no fixture), expandable fixtures list.
- Click a component row → preview (or toggle if it has fixtures); click
a fixture row → preview that fixture.
- Hover action buttons: `▷` preview + `</>` go-to-source on every row.
- "Go to source" opens the component file for component rows; for
fixture rows it opens the fixture file and jumps the cursor to the named
fixture key.
## Test coverage
- Runtime: DebugEventBus unit tests (8), handleDebugEvents endpoint (5),
fetch-router routes (4), integration across Express/Hono/Node/Fetch (9).
- Hooks: scanner + 16 fixture bundle-smoke test, regression guard
against `node_<builtin>` self-references, CSS collector test, stub-based
capture E2E, cross-kind controls remount, FormRenderer defensive
rendering.
- Inspector + webview: DebugStream reconnect (10), inspector components
(17), colors (9).
- Total: **178 tests** passing for the vscode-extension package; runtime
suite unchanged.
## Test plan
- [ ] `pnpm nx run copilotkit-vscode-extension:build` and `pnpm nx run
copilotkit-vscode-extension:test` both green
- [ ] F5 launches the Extension Dev Host with `test-workspace` open
- [ ] Hooks sidebar lists every fixture hook; click a row → preview
opens; `</>` button opens the source
- [ ] `▶️ Preview Component` CodeLens shows above every render hook in a
`.tsx` file; clicking it opens the preview
- [ ] Form controls drive the render live; cross-kind hook switches
(action ↔ custom-message) don't crash; a forced render-prop throw
recovers when a different hook is picked
- [ ] Imported-render fixtures (`ImportedAirQuality`,
`ImportedPollenReport`) bundle and preview correctly
- [ ] A2UI Catalog sidebar is the new webview, refresh works, `</>` on a
fixture opens the fixture file and reveals the named key
- [ ] AG-UI Inspector connects to `GET /debug-events`, filters + detail
work, events survive sidebar/panel switch, invalid URL shows red error
The dojo app was missing items under the langgraph column because
shell-dojo shipped a stale committed registry.json. The generator
only wrote to shell/, the dojo Dockerfile didn't run the generator
at build, and the CI path filter didn't rebuild the dojo when
manifest files changed.
Fix: emit from generate-registry.ts to shell, shell-dojo, and
shell-docs; add the generator step to shell-dojo's Dockerfile;
expand the deploy workflow's path filter to include packages/**
and shared/**; and refresh the committed registry/demo-content
JSON so files on disk match what the generator produces today.
The shell-dashboard app baked http://localhost:3000 into every demo and code link because NEXT_PUBLIC_SHELL_URL was never provided at build time and the source defaulted to localhost. Next.js inlines NEXT_PUBLIC_* at next build, so setting the value on Railway at runtime does nothing.
Fix: remove the silent localhost fallback, pass NEXT_PUBLIC_SHELL_URL as a Docker build arg from showcase_deploy.yml, and fail loudly if it's unset at build so this can't regress silently.
Adds a GitHub Actions workflow that curls /api/health on each of the
17 showcase starter services every 5 minutes to keep the agent's
in-memory state warm (JIT, module cache, connection pools). Railway
Pro tier doesn't sleep containers, but warm-state decays over idle.
Intentionally minimal: no Slack, no metrics, no state. Observability
is already covered by showcase_smoke-monitor.yml — this is strictly
keep-alive.
INTERIM — will be superseded by showcase-ops smoke probe cadence
once that service lands (see Notion proposal §2a).
## Summary
- Docs Sync workflow has been failing since 2026-04-21 with `fatal:
pathspec 'showcase/shell/src/content/' did not match any files` (exit
128) on the "Create PR for docs sync" step (example: run 24726678848).
- The path was moved by the shell platform restructure (#4109, #4112) —
the docs content now lives under `showcase/shell-docs/src/content/`, and
the sync marker at `showcase/shell-docs/.docs-sync-sha`. The sync script
(`showcase/scripts/sync-docs-from-main.ts`) already writes to those
locations; only the workflow's `git add` was still referencing the old
monolithic `showcase/shell/` path.
- One-line path fix.
## Test plan
- [ ] Verify yaml parses.
- [ ] Confirm next docs-sync run succeeds end-to-end (opens a sync PR
for the 11 pending review items from run 24726678848).
Regression from the 2026-04-21 incident: 18 production Railway services
were found with malformed image refs of the form
`ghcr.io/copilotkit/showcase-<slug>atest` (missing the `:` before
`latest`, so Docker treats `...atest` as the tag). Root cause was an
out-of-band MCP/manual mutation — no committed code touched those refs,
so the data has been fixed but no source-controlled guardrail exists.
Add a standalone script that queries Railway's GraphQL API for every
service in the CopilotKit Showcase project and asserts each image ref
matches the canonical shape `ghcr.io/copilotkit/<service-name>:latest`.
Wire it into showcase_deploy.yml as a pre-build job so any drift aborts
the workflow before the build matrix fans out.
On violation the script prints the service name, the current image, the
expected shape, and the reason, so the fix is obvious in the run log.
Slack classification in the notify job distinguishes a drift failure
from other pre-build failures.
Verified locally: 41 services pass against current Railway state; the
exported `validateImage` function rejects the exact `...atest`
corruption, mismatched service/image names, missing tags, wrong
registries, wrong tag values, and null sources (9/9 simulated cases).
The "Showcase: Docs Sync" workflow has been failing with
`fatal: pathspec 'showcase/shell/src/content/' did not match any files`
since the shell platform restructure (#4109, #4112) split the monolithic
`showcase/shell/` package into `showcase/shell-docs/`,
`showcase/shell-dashboard/`, and `showcase/shell-dojo/`.
The sync-docs-from-main.ts script already writes to
`showcase/shell-docs/src/content/` and maintains the sync marker at
`showcase/shell-docs/.docs-sync-sha`. Update the "Create PR for docs
sync" step to stage from the new location so the workflow can commit
synced content again.
matrix.service.timeout is already a number in the generated matrix
JSON, so wrapping it in fromJSON() was a no-op that only obscured
the expression. Drop the wrapper.
The HTTP_CODE="000" sentinel assigned immediately before the probe
loop is dead: the loop's first iteration unconditionally overwrites
HTTP_CODE before any reader runs. Remove it to avoid implying a
meaningful default where there isn't one.
The shell-family starters bundle shared demo content and tooling at
build time via showcase/scripts/bundle-demo-content.ts, which walks
showcase/shared and showcase/packages/*/manifest.yaml. A change to
any of those inputs can alter the generated bundle without touching
the package directory, so path-filter was under-reporting changes
and skipping deploys that actually needed to rebuild. Extend the
filters to shared/**, scripts/**, and packages/*/manifest.yaml so
those inputs trigger the correct downstream deploys.
Jordan Ritter