@odla-ai/cli 0.25.11 → 0.25.17

package/dist/index.d.cts

@@ -247,10 +247,10 @@ declare function adminAi(options: AdminAiOptions): Promise<void>;
 /** Stable, machine-readable division of work between the CLI, coding agent,
  * human operator, and Studio. Printed by `odla-ai capabilities --json`. */
 declare const CAPABILITIES: {
-    readonly cli: readonly ["start an email-bound device request without accepting a password or user session token", "register the app and enable configured services per environment", "issue, persist, and explicitly rotate configured service credentials", "add or remove app co-owners, each of whom self-provisions their own credentials for the shared db without any secret handoff", "write local Worker values and push deployed Worker secrets through Wrangler stdin", "store tenant-vault secrets (including the Clerk webhook secret and reserved Clerk secret key) write-only from stdin or an env var", "compose declared app-capability schema/rules, create guarded seeds when absent, and configure platform AI, auth, and deployment links", "apply Google Calendar booking config, then drive state-bound consent, status, discovery, and disconnect flows (bookings run live through the platform proxy)", "validate integration contracts offline and smoke-test a provisioned db environment plus anonymous capability routes", "read one versioned o11y status envelope spanning application RED, current live-sync freshness/load, the protected commit-to-visible canary, collector ingest/scheduler trust, provider-owned runtime metrics, and a bounded machine verdict", "inspect durable agent wakeups as a versioned JSON envelope and explicitly requeue one dead-lettered job with a scoped environment credential", "run app-attributed hosted security discovery and independent validation without provider keys", "connect/revoke source-read-only GitHub sources and drive commit-pinned hosted security jobs without PATs or provider keys", "let admins manage system AI policy and credentials, inspect attributed usage, and read immutable admin changes through separate short-lived capability-only approvals"];
+    readonly cli: readonly ["start an email-bound device request without accepting a password or user session token", "register the app and enable configured services per environment", "issue, persist, and explicitly rotate configured service credentials", "add or remove app co-owners, each of whom self-provisions their own credentials for the shared db without any secret handoff", "write local Worker values and push deployed Worker secrets through Wrangler stdin", "store tenant-vault secrets (including the Clerk webhook secret and reserved Clerk secret key) write-only from stdin or an env var", "compose declared app-capability schema/rules, create guarded seeds when absent, and configure platform AI, auth, and deployment links", "apply Google Calendar booking config, then drive state-bound consent, status, discovery, and disconnect flows (bookings run live through the platform proxy)", "validate integration contracts offline and smoke-test a provisioned db environment plus anonymous capability routes", "save and explicitly select non-secret named operator contexts with isolated credential caches; resolve and explain platform, app, environment, and credential provenance without authenticating; then run PM, Discussions, o11y, runbook, and identity operations outside a project checkout", "read one versioned o11y status envelope spanning application RED, exact Worker versions and Cloudflare colos observed in traffic, current live-sync freshness/load, the protected commit-to-visible canary, collector ingest/scheduler trust, provider-owned runtime metrics, account-scoped Durable Object, D1, and R2 evidence under odla-db, and a bounded machine verdict", "inspect durable agent wakeups as a versioned JSON envelope and explicitly requeue one dead-lettered job with a scoped environment credential", "run app-attributed hosted security discovery and independent validation without provider keys", "connect/revoke source-read-only GitHub sources and drive commit-pinned hosted security jobs without PATs or provider keys", "let admins manage system AI policy and credentials, inspect attributed usage, and read immutable admin changes through separate short-lived capability-only approvals"];
     readonly agent: readonly ["install and import the selected odla SDKs", "wrap the Worker with withObservability and choose useful telemetry", "install capability packages, mount their runtime routes, and make application-specific schema, rules, auth, UI, and migration decisions", "wire @odla-ai/calendar into trusted Worker code and keep the app admin key out of browsers"];
     readonly human: readonly ["provide the existing odla account email, then sign in and explicitly review/approve the exact device code", "review mirrored calendar/attendee disclosure and complete the server-issued Google consent flow", "consent to changes against an app's live database with --yes", "request destructive credential rotation explicitly", "review application semantics, security findings, releases, and merges", "approve redacted-source disclosure before hosted security reasoning", "approve GitHub App repository access separately from redacted-snippet disclosure"];
-    readonly studio: readonly ["view application telemetry, reconciled live-sync load/freshness, commit-to-visible canary health, provider-owned Worker runtime metrics, and environment state", "let signed-in users inventory/revoke their own agent grants and admins audit/global-revoke them", "review calendar connection, granted read scope, selected calendars, and sync health without exposing provider tokens", "perform manual credential recovery — for the primary owner or any co-owner — when the CLI's local shown-once copy is unavailable", "configure system AI purposes and view app/environment/run-attributed usage", "connect/revoke GitHub security sources and inspect ref-to-SHA jobs, routes, retention, coverage, and normalized reports"];
+    readonly studio: readonly ["view application telemetry grouped by exact Worker version, reconciled live-sync load/freshness, commit-to-visible canary health, provider-owned Worker runtime metrics, and environment state", "let signed-in users inventory/revoke their own agent grants and admins audit/global-revoke them", "review calendar connection, granted read scope, selected calendars, and sync health without exposing provider tokens", "perform manual credential recovery — for the primary owner or any co-owner — when the CLI's local shown-once copy is unavailable", "configure system AI purposes and view app/environment/run-attributed usage", "connect/revoke GitHub security sources and inspect ref-to-SHA jobs, routes, retention, coverage, and normalized reports"];
 };
 type Output = Pick<typeof console, "log">;
 /** Print the stable responsibility boundary for humans or agent tooling. */

package/dist/index.d.ts

@@ -247,10 +247,10 @@ declare function adminAi(options: AdminAiOptions): Promise<void>;
 /** Stable, machine-readable division of work between the CLI, coding agent,
  * human operator, and Studio. Printed by `odla-ai capabilities --json`. */
 declare const CAPABILITIES: {
-    readonly cli: readonly ["start an email-bound device request without accepting a password or user session token", "register the app and enable configured services per environment", "issue, persist, and explicitly rotate configured service credentials", "add or remove app co-owners, each of whom self-provisions their own credentials for the shared db without any secret handoff", "write local Worker values and push deployed Worker secrets through Wrangler stdin", "store tenant-vault secrets (including the Clerk webhook secret and reserved Clerk secret key) write-only from stdin or an env var", "compose declared app-capability schema/rules, create guarded seeds when absent, and configure platform AI, auth, and deployment links", "apply Google Calendar booking config, then drive state-bound consent, status, discovery, and disconnect flows (bookings run live through the platform proxy)", "validate integration contracts offline and smoke-test a provisioned db environment plus anonymous capability routes", "read one versioned o11y status envelope spanning application RED, current live-sync freshness/load, the protected commit-to-visible canary, collector ingest/scheduler trust, provider-owned runtime metrics, and a bounded machine verdict", "inspect durable agent wakeups as a versioned JSON envelope and explicitly requeue one dead-lettered job with a scoped environment credential", "run app-attributed hosted security discovery and independent validation without provider keys", "connect/revoke source-read-only GitHub sources and drive commit-pinned hosted security jobs without PATs or provider keys", "let admins manage system AI policy and credentials, inspect attributed usage, and read immutable admin changes through separate short-lived capability-only approvals"];
+    readonly cli: readonly ["start an email-bound device request without accepting a password or user session token", "register the app and enable configured services per environment", "issue, persist, and explicitly rotate configured service credentials", "add or remove app co-owners, each of whom self-provisions their own credentials for the shared db without any secret handoff", "write local Worker values and push deployed Worker secrets through Wrangler stdin", "store tenant-vault secrets (including the Clerk webhook secret and reserved Clerk secret key) write-only from stdin or an env var", "compose declared app-capability schema/rules, create guarded seeds when absent, and configure platform AI, auth, and deployment links", "apply Google Calendar booking config, then drive state-bound consent, status, discovery, and disconnect flows (bookings run live through the platform proxy)", "validate integration contracts offline and smoke-test a provisioned db environment plus anonymous capability routes", "save and explicitly select non-secret named operator contexts with isolated credential caches; resolve and explain platform, app, environment, and credential provenance without authenticating; then run PM, Discussions, o11y, runbook, and identity operations outside a project checkout", "read one versioned o11y status envelope spanning application RED, exact Worker versions and Cloudflare colos observed in traffic, current live-sync freshness/load, the protected commit-to-visible canary, collector ingest/scheduler trust, provider-owned runtime metrics, account-scoped Durable Object, D1, and R2 evidence under odla-db, and a bounded machine verdict", "inspect durable agent wakeups as a versioned JSON envelope and explicitly requeue one dead-lettered job with a scoped environment credential", "run app-attributed hosted security discovery and independent validation without provider keys", "connect/revoke source-read-only GitHub sources and drive commit-pinned hosted security jobs without PATs or provider keys", "let admins manage system AI policy and credentials, inspect attributed usage, and read immutable admin changes through separate short-lived capability-only approvals"];
     readonly agent: readonly ["install and import the selected odla SDKs", "wrap the Worker with withObservability and choose useful telemetry", "install capability packages, mount their runtime routes, and make application-specific schema, rules, auth, UI, and migration decisions", "wire @odla-ai/calendar into trusted Worker code and keep the app admin key out of browsers"];
     readonly human: readonly ["provide the existing odla account email, then sign in and explicitly review/approve the exact device code", "review mirrored calendar/attendee disclosure and complete the server-issued Google consent flow", "consent to changes against an app's live database with --yes", "request destructive credential rotation explicitly", "review application semantics, security findings, releases, and merges", "approve redacted-source disclosure before hosted security reasoning", "approve GitHub App repository access separately from redacted-snippet disclosure"];
-    readonly studio: readonly ["view application telemetry, reconciled live-sync load/freshness, commit-to-visible canary health, provider-owned Worker runtime metrics, and environment state", "let signed-in users inventory/revoke their own agent grants and admins audit/global-revoke them", "review calendar connection, granted read scope, selected calendars, and sync health without exposing provider tokens", "perform manual credential recovery — for the primary owner or any co-owner — when the CLI's local shown-once copy is unavailable", "configure system AI purposes and view app/environment/run-attributed usage", "connect/revoke GitHub security sources and inspect ref-to-SHA jobs, routes, retention, coverage, and normalized reports"];
+    readonly studio: readonly ["view application telemetry grouped by exact Worker version, reconciled live-sync load/freshness, commit-to-visible canary health, provider-owned Worker runtime metrics, and environment state", "let signed-in users inventory/revoke their own agent grants and admins audit/global-revoke them", "review calendar connection, granted read scope, selected calendars, and sync health without exposing provider tokens", "perform manual credential recovery — for the primary owner or any co-owner — when the CLI's local shown-once copy is unavailable", "configure system AI purposes and view app/environment/run-attributed usage", "connect/revoke GitHub security sources and inspect ref-to-SHA jobs, routes, retention, coverage, and normalized reports"];
 };
 type Output = Pick<typeof console, "log">;
 /** Print the stable responsibility boundary for humans or agent tooling. */

package/dist/index.js

@@ -49,7 +49,7 @@ import {
   startHostedSecurityJob,
   surfacePaths,
   validateInvocation
-} from "./chunk-FYSV7POZ.js";
+} from "./chunk-5MRY3V2K.js";
 export {
   AGENT_HARNESSES,
   CAPABILITIES,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@odla-ai/cli",
-  "version": "0.25.11",
+  "version": "0.25.17",
   "description": "Agent-operable CLI for odla provisioning, calendar consent and sync lifecycle, System AI administration, Worker secrets, security jobs, and smoke checks.",
   "license": "MIT",
   "homepage": "https://odla.ai/docs/packages/cli",

package/skills/odla-o11y-debug/SKILL.md

@@ -19,6 +19,24 @@ The core move is **cutting high-dimensional data to isolate the bad slice** —
 "is p95 high" but "*which* route / status / model / user is dragging it, and what
 do those requests share."
 
+Start with the stable aggregate envelope before composing individual routes:
+
+```cmd
+npx @odla-ai/cli context show --app <appId> --env prod --json
+npx @odla-ai/cli o11y status --app <appId> --env prod --minutes 60 --json
+```
+
+Both commands work outside a project checkout. `context show` reports where
+platform, app, environment, and developer-token context came from without
+printing a credential or starting a handshake. Prefer `ODLA_DEV_TOKEN` for an
+unattended agent; it is an ordinary revocable owner grant, not the collector's
+machine token.
+
+Schema v9 includes `applicationVersions`, which attributes request, error, and
+latency evidence to exact Worker version ids. Multiple values can be a rollout;
+an older id receiving traffic after the rollout is evidence, not a timestamp
+guess.
+
 ## Access
 
 All reads go through the platform shell, operator-authed, so the collector's token
@@ -29,13 +47,12 @@ GET https://odla.ai/o11y/<appId>/<endpoint>?env=prod
 Authorization: Bearer <developer-token>
 ```
 
-Get `<developer-token>` from `@odla-ai/cli` or the device handshake: start with
-`POST https://odla.ai/handshake` and
-`{"email":"<existing-odla-account>","label":"o11y debugger"}`, show the returned
-verification URL, then poll `POST /handshake/poll`. Email is a non-secret
-identifier; never ask for a password or session token. The matching account
-must sign in and explicitly review and approve the exact code. `env` is `prod`
-(default) or `dev`. If a read returns `{"error":"ae_not_configured"}` (501), the app's metrics
+Get `<developer-token>` through `@odla-ai/cli`: reuse `ODLA_DEV_TOKEN`, its
+private cache, or let the CLI start the email-bound device handshake. Email is
+a non-secret identifier; never ask for a password or session token. The
+matching account must sign in and explicitly review and approve the exact
+code. `env` is `prod` (default) or `dev`. If a read returns
+`{"error":"ae_not_configured"}` (501), the app's metrics
 reads aren't turned on yet — errors/traces still work; say so and fall back to them.
 
 ## Endpoints
@@ -49,12 +66,13 @@ reads aren't turned on yet — errors/traces still work; say so and fall back to
 | `GET /o11y/<app>/metrics/red?minutes=60` | request/error/latency by route |
 | `GET /o11y/<app>/metrics/timeseries?minutes=60` | RED over time (spot the spike) |
 | `GET /o11y/<app>/explore?dimension=&metric=` | **RED grouped by ANY dimension** |
+| `GET /o11y/odla-db/provider/cloudflare/capacity` | redacted account-scoped DO, D1, and R2 storage/load/latency evidence plus truncation |
 | `GET /o11y/<app>/llm/cost?minutes=1440` | spend by provider/model |
 | `GET /o11y/<app>/llm/cost/timeseries` | spend over time |
 | `GET /o11y/<app>/alerts` | configured triggers + firing history |
 
 `explore` dimensions: `route`, `status`, `name`, `kind`, `http_status`,
-`http_method`, `env`. `metric` (what to rank by): `p95_ms`, `avg_ms`, `requests`,
+`http_method`, `env`, `version`, `colo`. `metric` (what to rank by): `p95_ms`, `avg_ms`, `requests`,
 `errors`.
 
 ## Triage loops
@@ -69,11 +87,13 @@ downstream call (db, LLM, fetch) failed or was slow. Name the failing line.
 loop:
 1. `GET /metrics/timeseries?minutes=<window>` — confirm the spike and its window.
 2. `GET /explore?dimension=route&metric=p95_ms` — which route owns the p95? Then
-   pivot the dimension: `status`, `http_status`, `http_method`, `env` — keep
+   pivot the dimension: `status`, `http_status`, `http_method`, `env`, `version`,
+   `colo` — keep
    slicing until one value dominates the metric.
 3. Pull a couple of the offending traces (`/events` filtered to that route, then
    `/traces/<id>`) and read their spans + attributes. State **what the bad
-   requests share** (a route, a status code, a slow db/LLM span, a deploy env).
+   requests share** (a route, a status code, a slow db/LLM span, a deploy env,
+   or an exact Worker version).
 
 **Cost is climbing.** `GET /llm/cost/timeseries` (when) + `GET /llm/cost` (which
 model). A single expensive model or a runaway call count is usually the story.