From 7583bc1fbbad0c4c16dfd21ca36825788dc92969 Mon Sep 17 00:00:00 2001 From: Jinwoo-H Date: Wed, 12 Aug 2026 12:21:06 -0400 Subject: [PATCH] docs(skills): record remaining release evidence --- docs/agent-skill-sharing-implementation-checklist.md | 10 ++++++++-- docs/agent-skill-sharing-installation-plan.md | 6 ++++++ 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/docs/agent-skill-sharing-implementation-checklist.md b/docs/agent-skill-sharing-implementation-checklist.md index fe5b3df56da..9102440538e 100644 --- a/docs/agent-skill-sharing-implementation-checklist.md +++ b/docs/agent-skill-sharing-implementation-checklist.md @@ -672,8 +672,12 @@ does not mean the surrounding phase is complete. pipeline. - [x] Confirm or create a separate staging GCP project before exercising new lifecycle and IAM behavior. -- [ ] Confirm production names and quotas, including availability of - `onorca-cloud-skill-packages`. +- [x] Confirm production names and quotas, including availability of + `onorca-cloud-skill-packages`. A read-only 2026-08-12 check returned `404` for the global + bucket name, found no production `orca_skills` database or skill-database secret collision, + and confirmed the intended existing API service and runtime identity. Cloud Run exposes a + 1,000-job regional quota and 180 job runs per minute, while V1 reuses the existing API at a + 20-instance ceiling and existing Cloud SQL instance; no quota increase is required. - [x] Confirm `US` storage satisfies initial residency requirements. - [x] Confirm Cloud SQL connection and PostgreSQL user provisioning conventions. - [x] Keep Firestore, Cloud Tasks, and new Pub/Sub dependencies out of V1. @@ -1098,6 +1102,8 @@ source snapshot, dependencies, and package-manager store were removed after the transaction preserves the old bytes and returns retryable `skill-install-filesystem-failed`; the real Windows source remains intact for retry. - [ ] Terminate the runtime before and after each journal boundary and verify startup recovery. + The real-process macOS harness passes all 17 extraction, install, and removal crash points; + a fresh physical Windows/WSL run remains open. - [ ] Test permission-denied, read-only, disk-full, cancellation, runtime disconnect, and partial provider-coverage paths. Deterministic `EACCES` and `ENOSPC` transaction injection now proves the prior installed version remains intact; physical read-only/disk-full and disconnect diff --git a/docs/agent-skill-sharing-installation-plan.md b/docs/agent-skill-sharing-installation-plan.md index cf38bece355..078366d6d02 100644 --- a/docs/agent-skill-sharing-installation-plan.md +++ b/docs/agent-skill-sharing-installation-plan.md @@ -134,6 +134,12 @@ completed successfully and emitted only aggregate published and quarantine measu zero overdue objects and no identifiers or paths. SQL remained `NEVER`/`STOPPED`, all three Relay MIG targets remained zero, and production remained disabled and untouched. +A separate read-only production prerequisite check confirmed that +`onorca-cloud-skill-packages` is currently unallocated, the production skill database and secret +names do not collide, and the intended API service and runtime identity already exist. Cloud Run +quota headroom exceeds the existing 20-instance API ceiling; V1 adds no database instance and +requires no quota increase. No production resource was planned or changed. + Auth deploy run `31535179937` promoted `orca-cloud-auth-staging-00017-dug` with exact staging-only GitHub OIDC constraints, healthy JWKS, and zero deployment-window errors. API run `31535438327` promoted `orca-cloud-api-staging-00042-hef` with `authenticatedSmoke: true` and `skillSmoke: true`.