chore(infra): prep proposal-system for seahaven-prod deployment (#227)

* chore(infra): retarget prod to seahaven-prod account + OIDC deploy-role artifacts Retarget the CDK prod env from mgmt (328440206208, now frozen for workloads) to the dedicated seahaven-prod workload account (011934824531). proposal-system is the org's first prod tenant. Hard-block env=staging (still targets frozen mgmt) in resolveConfig until it is retargeted to seahaven-dev (710827005802). Add a WARN-only out-of-pipeline deploy guard in bin/app.ts. Add infra/deploy-role/: OIDC trust policy (sub scoped to Sea-Haven-Industries/proposal-system:ref:refs/heads/main), least-privilege permissions policy (AssumeRole on the verified cdk-hnb659fds bootstrap roles, deterministic site bucket, account-scoped CloudFront invalidation), and an idempotent creation script. Verified against live prod: bootstrap qualifier hnb659fds v32, OIDC provider present. Passed GPT-4.1 cross-review (APPROVE) and workflow red-team (CLEAN). Role NOT yet created — gated on /sh-security-review + the deploy go-ahead. Docs: README + CLAUDE.md reflect the prod account and pipeline-only deploy. * chore(infra): region-bound deploy-role DescribeStacks to us-east-1 (sh-security-review IAM-L2) * feat(infra): Aurora prod backup retention 14d + window; prod-only CDK context Bump Aurora automated-backup (PITR) retention 7->14d and set a preferred backup window for the prod tenant. Dedicated AWS Backup vault + cross-account restore test is a tracked follow-up (no org central-backup design exists yet). Prune the stale mgmt-account AZ context; prod (011934824531) is the only deploy target.
2026-09-30 04:13:13 +00:00 · 2026-07-15 14:44:35 -04:00 · 2026-07-15 14:44:35 -04:00 · 0f9d27fbf0
commit 0f9d27fbf0
parent cac4384f0d
10 changed files with 260 additions and 17 deletions
--- a/CLAUDE.md
+++ b/CLAUDE.md
@ -35,7 +35,7 @@ Internal: Lambdas → .NET Function URL with Secrets Manager API key via custom

 ## Infrastructure

-AWS us-east-1, Aurora PostgreSQL 15 Serverless v2 (RDS Data API + pgvector), S3, SQS+DLQ, Cognito+Google OAuth, Bedrock KB (Aurora pgvector store — ADR 0001), GitHub Actions OIDC, CloudFront+S3 OAC
+Deploys to the **seahaven-prod** account (`011934824531`), us-east-1. (mgmt `328440206208` is frozen for workloads; staging is hard-disabled in `infra/lib/config.ts` until retargeted to seahaven-dev `710827005802`.) Aurora PostgreSQL 15 Serverless v2 (RDS Data API + pgvector), S3, SQS+DLQ, Cognito+Google OAuth, Bedrock KB (Aurora pgvector store — ADR 0001), GitHub Actions OIDC, CloudFront+S3 OAC. Prod deploys run via the `deploy.yaml` pipeline (workflow_dispatch) only — no local `cdk deploy` to prod.

 ## Agent Delegation Rules

--- a/README.md
+++ b/README.md
@ -47,7 +47,11 @@ proposal-system/

 ## AWS Resources

-All resources are in **us-east-1** (account 328440206208).
+All resources deploy to **us-east-1** in the **seahaven-prod** workload account (`011934824531`).
+The mgmt account (`328440206208`) is frozen for new workloads. Staging is not currently
+deployable — it still references the frozen mgmt account and is hard-blocked in `infra/lib/config.ts`
+until retargeted to seahaven-dev (`710827005802`). Prod deploys go through the `deploy.yaml`
+pipeline (workflow_dispatch) only; no manual/local `cdk deploy` to prod.

 | CDK Stack | Key Resources |
 |---|---|
--- a/infra/bin/app.ts
+++ b/infra/bin/app.ts
@ -6,11 +6,24 @@ import { resolveConfig } from '../lib/config';

 const app = new cdk.App();

-// Multi-env (PR2): `-c env=staging` (default prod). prod keeps the exact construct IDs
-// and stack names of the deployed stacks (stackSuffix=''); only staging is suffixed.
+// Multi-env (PR2): `-c env=staging` (default prod). prod targets seahaven-prod
+// (011934824531) with no stack-name suffix; staging is suffixed and currently hard-disabled
+// (see resolveConfig — mgmt account is frozen).
 const config = resolveConfig(app);
 const { env, stackSuffix } = config;

+// Pipeline-only deploy signal (WARN, not block). Prod deploys must go through the cd-cdk
+// pipeline (GitHub Actions sets CI/GITHUB_ACTIONS). A local synth/deploy to prod is a
+// drift risk + "no manual prod deploys" violation. True enforcement is an org-baseline SCP
+// (tracked follow-up); this is the cheap in-repo backstop.
+if (config.envName === 'prod' && !process.env.CI && !process.env.GITHUB_ACTIONS) {
+  // eslint-disable-next-line no-console
+  console.warn(
+    '\n⚠️  Running the PROD CDK app outside CI. Prod deploys must go through the cd-cdk ' +
+      'pipeline (deploy.yaml, workflow_dispatch). Do NOT `cdk deploy` to prod locally.\n',
+  );
+}
+
 const foundation = new FoundationStack(app, `proposal-system-foundation${stackSuffix}`, {
  stackName: `proposal-system-foundation${stackSuffix}`,
  env,
--- a/infra/cdk.context.json
+++ b/infra/cdk.context.json
@ -1,5 +1,5 @@
 {
-  "availability-zones:account=328440206208:region=us-east-1": [
+  "availability-zones:account=011934824531:region=us-east-1": [
    "us-east-1a",
    "us-east-1b",
    "us-east-1c",
--- a/infra/deploy-role/README.md
+++ b/infra/deploy-role/README.md
@ -0,0 +1,64 @@
+# seahaven-prod OIDC deploy role — proposal-system
+
+IAM artifacts for the GitHub Actions OIDC deploy role that lets the
+`Sea-Haven-Industries/proposal-system` repo deploy the CDK stacks and sync the
+frontend site bucket into **seahaven-prod**. Artifacts only — nothing here has
+been applied to AWS.
+
+## Resolved facts (Phase 0, read-only verification)
+
+| Field | Value |
+|---|---|
+| Account | `011934824531` (seahaven-prod) |
+| Region | `us-east-1` |
+| CDK qualifier | `hnb659fds` (AWS CDK **default** — bootstrap v32; no custom synthesizer needed) |
+| OIDC provider ARN | `arn:aws:iam::011934824531:oidc-provider/token.actions.githubusercontent.com` (EXISTS) |
+| Role name | `githubdeploy-proposal-system` |
+| Subject scope | `repo:Sea-Haven-Industries/proposal-system:ref:refs/heads/main` (exact, no wildcard) |
+| Site bucket | `proposal-system-web-011934824531` (frontend-stack convention) |
+
+## Files
+
+- **trust-policy.json** — Web-identity trust policy. Federated principal is the
+  existing GitHub OIDC provider. `sts:AssumeRoleWithWebIdentity` gated by two
+  StringEquals conditions: `aud == sts.amazonaws.com` and an **exact** `sub`
+  match on the proposal-system repo's `main` branch (no `StringLike`, no
+  wildcard). Mirrors the structure of the existing
+  `githubdeploy-seahaven-org-baseline` role.
+
+- **permissions-policy.json** — Least-privilege inline policy:
+  - `sts:AssumeRole` on the four CDK bootstrap roles (deploy, file-publishing,
+    lookup, image-publishing) scoped to qualifier `hnb659fds`, account, and
+    region. This is how a CDK deploy actually gains its power — no direct
+    service permissions are granted to the deploy role itself.
+  - `cloudformation:DescribeStacks` on `*` for the `cdk deploy` /
+    change-set health check.
+  - `s3:PutObject`/`DeleteObject`/`ListBucket` on the site bucket and its
+    objects for the frontend asset sync.
+  - `cloudfront:CreateInvalidation` on `*`, constrained by
+    `aws:ResourceAccount == 011934824531` (distribution ARNs aren't known at
+    author time; the account condition prevents cross-account use).
+
+- **create-deploy-role.sh** — Idempotent bash (`aws --profile prod`). Verifies
+  the profile resolves to `011934824531`, then create-role (or
+  update-assume-role-policy if it exists) + put-role-policy. Safe to re-run.
+  **Gated:** do not execute until GPT-4.1 cross-review AND `/sh-security-review`
+  pass (IAM/trust change).
+
+## Applying (after gates pass)
+
+```bash
+./create-deploy-role.sh
+```
+
+Then point the GitHub Actions workflow's `aws-actions/configure-aws-credentials`
+step at the printed role ARN.
+
+## Placeholders / follow-ups
+
+- **None outstanding.** Qualifier resolved to the real value `hnb659fds`; no
+  `<QUALIFIER>` placeholder remains. OIDC provider exists, so no provider
+  creation prerequisite.
+- CloudFront invalidation is scoped by account, not by distribution ARN — tighten
+  to the specific distribution ARN once frontend-stack is deployed if you want
+  per-resource least privilege.
--- a/infra/deploy-role/create-deploy-role.sh
+++ b/infra/deploy-role/create-deploy-role.sh
@ -0,0 +1,71 @@
+#!/usr/bin/env bash
+###############################################################################
+# create-deploy-role.sh
+#
+# Creates / updates the GitHub Actions OIDC deploy role
+# `githubdeploy-proposal-system` in AWS account 011934824531 (seahaven-prod),
+# us-east-1, for the Sea-Haven-Industries/proposal-system repo (main branch).
+#
+# GATE — DO NOT EXECUTE until BOTH of the following have passed:
+#   1. GPT-4.1 cross-family review (IAM policy / trust-policy change), via:
+#        python3 ~/Documents/repositories/seahaven/security-review/cross_review.py \
+#          "Review this IAM deploy-role trust+permissions for over-permission: <artifacts>"
+#      (STATUS 2026-07-14: RUN — verdict APPROVE, no BLOCK.)
+#   2. /sh-security-review (deep agentic pass — IaC/IAM is a gated surface)
+#
+# This is an IAM/trust change: run BOTH gates and resolve every confirmed
+# critical/high before running. This script mutates AWS; the artifact-authoring
+# task did NOT run it. It is idempotent and safe to re-run.
+#
+# Resolved facts (Phase 0, read-only verification):
+#   Account   : 011934824531 (seahaven-prod)
+#   Region    : us-east-1
+#   Qualifier : hnb659fds  (AWS CDK DEFAULT — no custom synthesizer needed)
+#   OIDC prov : arn:aws:iam::011934824531:oidc-provider/token.actions.githubusercontent.com
+###############################################################################
+
+set -euo pipefail
+
+PROFILE="prod"
+ROLE_NAME="githubdeploy-proposal-system"
+POLICY_NAME="proposal-system-deploy"
+ACCOUNT_ID="011934824531"
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+TRUST_POLICY="file://${SCRIPT_DIR}/trust-policy.json"
+PERMS_POLICY="file://${SCRIPT_DIR}/permissions-policy.json"
+
+echo "==> Verifying active account for profile '${PROFILE}'..."
+CALLER_ACCOUNT="$(aws --profile "${PROFILE}" sts get-caller-identity --query Account --output text)"
+if [[ "${CALLER_ACCOUNT}" != "${ACCOUNT_ID}" ]]; then
+  echo "ERROR: profile '${PROFILE}' resolves to account ${CALLER_ACCOUNT}, expected ${ACCOUNT_ID}. Aborting." >&2
+  exit 1
+fi
+
+echo "==> Ensuring role '${ROLE_NAME}' exists with the correct trust policy..."
+if aws --profile "${PROFILE}" iam get-role --role-name "${ROLE_NAME}" >/dev/null 2>&1; then
+  echo "    Role exists — updating assume-role (trust) policy."
+  aws --profile "${PROFILE}" iam update-assume-role-policy \
+    --role-name "${ROLE_NAME}" \
+    --policy-document "${TRUST_POLICY}"
+else
+  echo "    Role absent — creating."
+  aws --profile "${PROFILE}" iam create-role \
+    --role-name "${ROLE_NAME}" \
+    --assume-role-policy-document "${TRUST_POLICY}" \
+    --description "GitHub Actions OIDC deploy role for Sea-Haven-Industries/proposal-system (main)" \
+    --max-session-duration 3600 \
+    --tags Key=project,Value=proposal-system Key=managed-by,Value=create-deploy-role.sh
+fi
+
+echo "==> Putting inline permissions policy '${POLICY_NAME}' (create-or-replace)..."
+aws --profile "${PROFILE}" iam put-role-policy \
+  --role-name "${ROLE_NAME}" \
+  --policy-name "${POLICY_NAME}" \
+  --policy-document "${PERMS_POLICY}"
+
+ROLE_ARN="$(aws --profile "${PROFILE}" iam get-role --role-name "${ROLE_NAME}" \
+  --query Role.Arn --output text)"
+
+echo "==> Done. Deploy role ready:"
+echo "    ${ROLE_ARN}"
+echo "    Configure the GitHub Actions workflow to assume this ARN via aws-actions/configure-aws-credentials."
--- a/infra/deploy-role/permissions-policy.json
+++ b/infra/deploy-role/permissions-policy.json
@ -0,0 +1,51 @@
+{
+    "Version": "2012-10-17",
+    "Statement": [
+        {
+            "Sid": "AssumeCdkBootstrapRoles",
+            "Effect": "Allow",
+            "Action": "sts:AssumeRole",
+            "Resource": [
+                "arn:aws:iam::011934824531:role/cdk-hnb659fds-deploy-role-011934824531-us-east-1",
+                "arn:aws:iam::011934824531:role/cdk-hnb659fds-file-publishing-role-011934824531-us-east-1",
+                "arn:aws:iam::011934824531:role/cdk-hnb659fds-lookup-role-011934824531-us-east-1",
+                "arn:aws:iam::011934824531:role/cdk-hnb659fds-image-publishing-role-011934824531-us-east-1"
+            ]
+        },
+        {
+            "Sid": "CdkDeployHealthCheck",
+            "Effect": "Allow",
+            "Action": "cloudformation:DescribeStacks",
+            "Resource": "*",
+            "Condition": {
+                "StringEquals": {
+                    "aws:RequestedRegion": "us-east-1"
+                }
+            }
+        },
+        {
+            "Sid": "FrontendSiteBucketSync",
+            "Effect": "Allow",
+            "Action": [
+                "s3:PutObject",
+                "s3:DeleteObject",
+                "s3:ListBucket"
+            ],
+            "Resource": [
+                "arn:aws:s3:::proposal-system-web-011934824531",
+                "arn:aws:s3:::proposal-system-web-011934824531/*"
+            ]
+        },
+        {
+            "Sid": "CloudFrontInvalidation",
+            "Effect": "Allow",
+            "Action": "cloudfront:CreateInvalidation",
+            "Resource": "*",
+            "Condition": {
+                "StringEquals": {
+                    "aws:ResourceAccount": "011934824531"
+                }
+            }
+        }
+    ]
+}
--- a/infra/deploy-role/trust-policy.json
+++ b/infra/deploy-role/trust-policy.json
@ -0,0 +1,18 @@
+{
+    "Version": "2012-10-17",
+    "Statement": [
+        {
+            "Effect": "Allow",
+            "Principal": {
+                "Federated": "arn:aws:iam::011934824531:oidc-provider/token.actions.githubusercontent.com"
+            },
+            "Action": "sts:AssumeRoleWithWebIdentity",
+            "Condition": {
+                "StringEquals": {
+                    "token.actions.githubusercontent.com:aud": "sts.amazonaws.com",
+                    "token.actions.githubusercontent.com:sub": "repo:Sea-Haven-Industries/proposal-system:ref:refs/heads/main"
+                }
+            }
+        }
+    ]
+}
--- a/infra/lib/config.ts
+++ b/infra/lib/config.ts
@ -1,10 +1,17 @@
 import * as cdk from 'aws-cdk-lib';

 // Multi-environment configuration (PR2). Resolved from CDK context: `-c env=staging`,
-// default `prod`. CRITICAL: the prod config MUST keep the exact construct IDs, stack
-// names, and resource settings the deployed stacks already use — renaming a deployed
-// CloudFormation stack triggers replace-and-delete, which would destroy the RDS instance.
-// Only non-prod environments take a stack-name suffix.
+// default `prod`. Only non-prod environments take a stack-name suffix.
+//
+// prod now targets the dedicated seahaven-prod workload account (011934824531); the mgmt
+// account (328440206208) is FROZEN for new workloads. The prior proposal-system footprint
+// was fully torn down 2026-07-14, so THIS first prod deploy is the last safe window to
+// rename construct IDs / stack names. After prod go-live, renaming a deployed stack (or a
+// stateful construct ID) triggers replace-and-delete, which would destroy the Aurora cluster.
+//
+// STAGING is intentionally NOT deployable: it still points at the frozen mgmt account, and
+// resolveConfig() hard-throws on `env=staging` until it is retargeted to seahaven-dev
+// (710827005802). See the guard in resolveConfig below.

 export type EnvName = 'prod' | 'staging';

@ -28,14 +35,15 @@ export interface EnvConfig {
  readonly retainData: boolean;
 }

-const ACCOUNT = '328440206208';
+// seahaven-prod workload account (org prod OU). mgmt 328440206208 is frozen for workloads.
+const PROD_ACCOUNT = '011934824531';
+// Frozen mgmt account — staging still references it and must NOT be deployed there.
+const MGMT_ACCOUNT_FROZEN = '328440206208';
 const REGION = 'us-east-1';

-// Prod values reproduce the currently-deployed stacks EXACTLY (verified against
-// foundation-stack.ts / compute-stack.ts) so `cdk diff` shows no prod changes.
 const PROD: EnvConfig = {
  envName: 'prod',
-  env: { account: ACCOUNT, region: REGION },
+  env: { account: PROD_ACCOUNT, region: REGION },
  stackSuffix: '',
  s3CorsOrigins: ['https://proposals.seahaven.com', 'http://localhost:5173'],
  apiCorsOrigins: [
@ -53,11 +61,13 @@ const PROD: EnvConfig = {
  retainData: true,
 };

-// Staging: same AWS account (Adam, 2026-06-12), suffixed stacks, no permanent domain
-// yet (CloudFront default URL + localhost), data not retained.
+// Staging: ⚠️ STILL points at the FROZEN mgmt account (MGMT_ACCOUNT_FROZEN). It is NOT
+// deployable — resolveConfig() throws on env=staging. Retarget to seahaven-dev
+// (710827005802) in a dedicated follow-up before ever enabling staging deploys.
+// Suffixed stacks, CloudFront default URL + localhost, data not retained.
 const STAGING: EnvConfig = {
  envName: 'staging',
-  env: { account: ACCOUNT, region: REGION },
+  env: { account: MGMT_ACCOUNT_FROZEN, region: REGION },
  stackSuffix: '-staging',
  s3CorsOrigins: ['http://localhost:5173'],
  apiCorsOrigins: ['http://localhost:5173'],
@ -76,5 +86,14 @@ export function resolveConfig(app: cdk.App): EnvConfig {
  if (!config) {
    throw new Error(`Unknown env '${name}'. Use -c env=prod (default) or -c env=staging.`);
  }
+  // Hard guard: staging still targets the frozen mgmt account (328440206208). Block any
+  // synth/deploy under env=staging until it is retargeted to seahaven-dev (710827005802),
+  // so an accidental `-c env=staging` deploy can never land in the frozen mgmt account.
+  if (name === 'staging') {
+    throw new Error(
+      'env=staging is disabled: it targets the FROZEN mgmt account (328440206208). ' +
+        'Retarget STAGING to seahaven-dev (710827005802) in config.ts before using it.',
+    );
+  }
  return config;
 }
--- a/infra/lib/foundation-stack.ts
+++ b/infra/lib/foundation-stack.ts
@ -105,7 +105,10 @@ export class FoundationStack extends cdk.Stack {
      serverlessV2MaxCapacity: 4,
      enableDataApi: true,
      storageEncrypted: true,
-      backup: { retention: cdk.Duration.days(7) },
+      // Aurora native automated backups (PITR). 14-day retention for the prod tenant;
+      // a dedicated AWS Backup vault + cross-account restore test is a tracked follow-up
+      // (no org central-backup design exists yet — see the prod-deploy plan Phase 4 fallback).
+      backup: { retention: cdk.Duration.days(14), preferredWindow: '07:00-08:00' },
      deletionProtection: config.retainData,
      removalPolicy: config.retainData ? cdk.RemovalPolicy.RETAIN : cdk.RemovalPolicy.DESTROY,
      defaultDatabaseName: 'proposals',