feat(appkit): surface caller expiry and support local OBO

[MarioCadenas]

Summary

Caller operations expose downstream credential rejection as IdentityExpiredError (IDENTITY_EXPIRED), and local apps can exercise real user execution through the normal npm run dev workflow.

This is layer 4 of the stack, based on #595 (execution-standalone-user). Design: design-docs/execution-identity-e2e.md, section 5.1.4. Cache partitioning is in #594.

• AppKit-managed spans record appkit.execution.principal, appkit.execution.principal_id, and the initiating user's appkit.execution.actor_id when present. Tokens are never span attributes.
• The server plugin injects forwarded user headers before plugin and custom routes in development. It uses DATABRICKS_TOKEN with DATABRICKS_HOST first, otherwise the explicitly configured DATABRICKS_CONFIG_PROFILE. No separate proxy command, target, or port is needed.
• Credential lookup and refresh are shared and kept in memory. Lookup and refresh failures return 401. Existing forwarded tokens are preserved. Injection accepts same-origin loopback requests and can be disabled with APPKIT_DEV_OBO=false.
• Header injection does not open a caller scope. asUser(req) selects user execution; unscoped calls keep the app's configured credentials. SP remains the ordinary default, group is deferred, and the marked DEV_OBO_FALLBACK remains when injection is disabled or unconfigured.
• Validate relative imports while packaging the CLI, with the validator dependency and Knip entry declared explicitly.

Local emulation supplies user credentials; it does not emulate platform consent, scope enforcement, or resource provisioning. A real SP-versus-user comparison requires the app's default credentials to belong to an SP.

Compatibility

Plugin execute() preserves its existing failed-result envelope and adds an optional typed error field. Throwing APIs expose IdentityExpiredError, and HTTP/SSE agent responses carry its stable code. The error retains only a token fingerprint, not credential-bearing upstream errors. Non-401 failures and SP execution retain their existing behavior. Existing public compatibility aliases remain available.

Verification

• pnpm -r typecheck, full monorepo unit suite: 5,301 passed, one existing skip.
• Original dev-playground OBO regression tests pass unchanged.
• pnpm build, pnpm docs:build, formatting/lint, and Knip.
• AppKit tarball build and packaged CLI --help smoke check.
• Coverage includes typed expiry, token-free errors, stream failures, span identity, token priority, explicit profile selection, refresh failure, header preservation, local request restrictions, and actual HTTP middleware injection.
• Tests use fake credentials and loopback servers. No live Databricks credentials were used.

[github-actions]

📦 Bundle size report

Compared against bundle-size-baseline.json (main).

@databricks/appkit

npm tarball (packed): 1.2 MB (+21 KB) — gzipped download (dist + bin; excludes release-only docs/NOTICE).

dist	raw	gzip
JS (runtime)	1.2 MB (+23 KB)	437 KB (+8.8 KB)
Type declarations	452 KB (+6.1 KB)	164 KB (+2.4 KB)
Source maps	2.4 MB (+41 KB)	818 KB (+16 KB)
Other	11 KB	3.7 KB
Total	4.1 MB (+70 KB)	1.4 MB (+27 KB)

Per-entry composition (own code — deps external (as shipped))

Entry	Initial (gz)	Lazy (gz)	Total (gz)	node_modules (min)	Own code (min)
.	99 KB (+2.9 KB)	2.5 KB (-2 B)	102 KB (+2.9 KB)	external	324 KB (+8.6 KB)
./beta	95 KB (+1.4 KB)	460 B (-18 B)	95 KB (+1.4 KB)	external	287 KB (+4.7 KB)
./testing	40 KB (+2.0 KB)	32 KB (+1.0 KB)	72 KB (+3.1 KB)	external	210 KB (+7.8 KB)
./tsdown	520 B	0 B	520 B	external	813 B
./type-generator	23 KB	0 B	23 KB	external	65 KB

Chunks:

Entry	Chunk	Load	Size (gz)
.	index.js	initial	95 KB
.	utils.js	initial	4.6 KB
.	remote-tunnel-manager.js	lazy	2.5 KB
./beta	beta.js	initial	78 KB
./beta	stream-manager.js	initial	5.9 KB
./beta	service-context.js	initial	4.1 KB
./beta	databricks.js	initial	3.3 KB
./beta	wide-event-emitter.js	initial	3.2 KB
./beta	client.js	initial	593 B
./beta	index.js	initial	20 B
./beta	supervisor-api.js	lazy	191 B
./beta	databricks.js	lazy	155 B
./beta	index.js	lazy	114 B
./testing	manifest.js	initial	27 KB
./testing	index.js	initial	10 KB
./testing	wide-event-emitter.js	initial	2.9 KB
./testing	index.js	lazy	28 KB
./testing	remote-tunnel-manager.js	lazy	2.5 KB
./testing	utils.js	lazy	1.8 KB
./tsdown	index.js	initial	520 B
./type-generator	index.js	initial	23 KB

@databricks/appkit-ui

npm tarball (packed): 350 KB — gzipped download (dist + bin; excludes release-only docs/NOTICE).

dist	raw	gzip
JS (runtime)	395 KB	132 KB
Type declarations	229 KB	84 KB
Source maps	766 KB	253 KB (+1 B)
CSS	16 KB	3.2 KB
Total	1.4 MB	472 KB (+1 B)

Per-entry composition (consumer bundle — deps bundled, peerDeps external)

Entry	Initial (gz)	Lazy (gz)	Total (gz)	node_modules (min)	Own code (min)
./js	5.3 KB	49 KB	55 KB	208 KB	14 KB
./js/beta	20 B	0 B	20 B	0 B	0 B
./react	432 KB	49 KB	481 KB	1.3 MB	177 KB
./react/beta	1.0 KB	0 B	1.0 KB	0 B	1.9 KB

Chunks:

Entry	Chunk	Load	Size (gz)
./js	index.js	initial	5.2 KB
./js	chunk	initial	120 B
./js	apache-arrow	lazy	49 KB
./js/beta	beta.js	initial	20 B
./react	index.js	initial	430 KB
./react	tslib	initial	2.1 KB
./react	apache-arrow	lazy	49 KB
./react/beta	beta.js	initial	1.0 KB

[github-actions]

🤖 AppKit PR bot

🔬 Run evals

Start an eval for this PR from the evals-monitor app: Go to Evals Monitor →

📦 Try this PR's app template

Scaffolds a new app from this PR's SDK build. Run it in any folder (requires the GitHub CLI — gh auth login — and the Databricks CLI):

gh run download 36878400847 -R databricks/appkit -n appkit-template-0.81.0-pr.7677039-execution-identity-lifecycle-597 -D appkit-pr-597 \
  && unzip -o "appkit-pr-597/appkit-template-0.81.0-pr.7677039-execution-identity-lifecycle-597.zip" -d "appkit-pr-597" \
  && databricks apps init --template "appkit-pr-597"

The template pins @databricks/appkit and @databricks/appkit-ui to tarballs built from this branch, so the scaffolded app runs against this PR's code.