feat(iam): add external-dev backend Terraform substrate (#131)

* feat(iam): add external-dev backend terraform substrate

* fix(iam): require boundaries for SHOC policy writes

.gitignore

@@ -1,5 +1,5 @@
 node_modules/
-cdk.out/
+cdk.out*/
 *.js
 *.d.ts
 *.js.map

README.md

@@ -18,7 +18,8 @@ like any other stack.
 > only) until 2026-07-14, when the external-dev member baseline was merged in
 > and the repo renamed. Deployed CloudFormation stack names are unchanged.
 
-Stacks (deployed by the CD workflow — one job per target account):
+Stacks (normally deployed by one CD job per target account; staged exceptions
+are noted):
 
 | Stack | Account | Region | Purpose |
 |---|---|---|---|
@@ -30,6 +31,7 @@ Stacks (deployed by the CD workflow — one job per target account):
 | `seahaven-backup-offsite` | 328440206208 | us-west-2 | Governance-locked offsite copy vault (C-7) |
 | `seahaven-org-governance` | 328440206208 | us-east-1 | AWS Organizations OU tree + SCPs (incl. the cdk-imported external-dev guardrails) |
 | `seahaven-external-dev-baseline` | 396287094661 | us-east-1 | Member-account baseline: Config, GuardDuty, Security Hub (FSBP + CIS v3.0), Access Analyzer, flow logs, budget |
+| `seahaven-terraform-substrate` | 396287094661 | us-east-1 | Staged manually until SHOC role imports complete; HCP roles/deploy boundaries for the backend rehearsal, referencing the existing OIDC provider |
 | `seahaven-security-baseline` | 001520130573 | us-east-1 | Member-account baseline for the delegated security-admin account (same construct set) |
 | `seahaven-dev-baseline` | 710827005802 | us-east-1 | Member-account baseline for internal dev/staging (org-managed detection — no local GuardDuty/SecurityHub) |
 | `seahaven-prod-baseline` | 011934824531 | us-east-1 | Member-account baseline for production workloads (org-managed detection; mgmt account frozen for new workloads) |
@@ -67,6 +69,7 @@ the TypeScript source — no separate compile step needed for `cdk synth` /
 | `prod-baseline` | `seahaven-prod-baseline` | 011934824531 | us-east-1 | `lib/member-baseline-stack.ts` (orgManagedDetection) |
 | `deploy-substrate-prod` | `seahaven-deploy-substrate` | 011934824531 | us-east-1 | `lib/deploy-substrate-stack.ts` |
 | `deploy-substrate-dev` | `seahaven-deploy-substrate` | 710827005802 | us-east-1 | `lib/deploy-substrate-stack.ts` |
+| `terraform-substrate-external-dev` | `seahaven-terraform-substrate` | 396287094661 | us-east-1 | `lib/terraform-substrate-stack.ts` |
 | `dynamodb-cmk-prod` | `seahaven-dynamodb-cmk` | 011934824531 | us-east-1 | `lib/dynamodb-cmk-stack.ts` |
 | `alarm-topic-prod` | `seahaven-alarm-topic` | 011934824531 | us-east-1 | `lib/alarm-topic-stack.ts` |
 | `app-web-acl-prod` | `seahaven-app-web-acl` | 011934824531 | us-east-1 | `lib/app-web-acl-stack.ts` |
@@ -214,9 +217,11 @@ created with the boundary already attached.
 
 `lib/terraform-substrate-stack.ts` + `lib/terraform-substrate/terraform-substrate.template.yaml`
 deploy `seahaven-terraform-substrate` into each member account that hosts
-Terraform-managed workloads (currently seahaven-prod and seahaven-dev; never
+Terraform-managed workloads (currently seahaven-prod, seahaven-dev, and
-mgmt — mgmt stays SAM until its stacks migrate out). It contains only the
+external-dev; never mgmt — mgmt stays SAM until its stacks migrate out).
-shared account-level plumbing:
+Prod/dev use the shared IAM-management policy. External-dev references its
+existing `app.terraform.io` provider and carries only exact SHOC
+import/adoption roles:
 
 - the `app.terraform.io` OIDC identity provider (audience
   `aws.workload.identity`; Retain — it is the federation anchor for every
@@ -226,7 +231,19 @@ shared account-level plumbing:
   `seahaven-lambda-execution-boundary` allow-list owned by the
   deploy-substrate stack — hence the explicit stack dependency in
   `bin/app.ts`) plus the `DenyBoundaryTampering` / `DenyBoundaryPolicyEdit`
-  / `DenySelfMutation` backstops.
+  / `DenySelfMutation` backstops,
+- external-dev-only deploy boundaries
+  `shoc-backend-{tf-poc,dev,staging}-deploy-boundary`. Each is the maximum
+  current policy for one exact `githubdeploy-shoc-backend-*` role. Dev
+  temporarily retains its live broad `elasticbeanstalk-*` S3 grants so
+  boundary attachment cannot regress deployment before the separately
+  reviewed policy narrowing,
+- external-dev-only runtime boundaries
+  `shoc-backend-{tf-poc,dev,staging}-runtime-boundary`. These retain only the
+  account-scoped S3, environment health/log, and X-Ray portions of
+  `AWSElasticBeanstalkWebTier`, plus each environment's exact secrets/KMS/STS
+  data plane. They deliberately exclude the managed policy's 2026
+  Bedrock/Marketplace additions.
 
 **This policy derives from `seahaven-cfn-exec-iam-management` but is
 deliberately stricter — it is not a mirror.** The 2026-07-30 security review
@@ -255,6 +272,110 @@ deliberately NOT pre-provisioned — they are appended to the template at each
 stack's migration time so an account never carries trust for workspaces that
 do not deploy to it.
 
+**External-dev SHOC role adoption is a staged CloudFormation import, not a
+normal first deploy.** Exactly four roles exist today:
+`hcptf-shoc-backend-{dev,staging}` and their `-plan` partners. The tf-poc pair
+does not exist. Two independent CDK contexts make each transition explicit:
+`enableShocBackendPocRoles` and `enableShocBackendLiveRoles`. Both are
+version-controlled as `false` in `cdk.json` for the initial rollout.
+`terraform-substrate-external-dev` is deliberately absent from the automatic
+external-dev deploy job during this sequence; `external-dev-baseline` remains
+automatic and unchanged.
+
+1. Create the role-free base stack:
+
+   ```bash
+   npx cdk deploy terraform-substrate-external-dev \
+     -c enableShocBackendPocRoles=false \
+     -c enableShocBackendLiveRoles=false
+   ```
+
+   `CreateOIDCProvider=false` is fixed in `bin/app.ts`; the external-dev
+   account therefore creates neither the existing provider, SHOC roles, nor
+   the prod/dev-only shared IAM policy. The base stack does create all six
+   retained external-dev deploy/runtime boundary policies.
+2. Set `enableShocBackendPocRoles` to `true` in `cdk.json`, leave the live gate
+   `false`, review the synthesized two-role addition, then run the normal
+   external-dev stack update. This creates only the new tf-poc HCP plan/apply
+   pair. The retained POC CDK stack references
+   `shoc-backend-tf-poc-deploy-boundary` when it creates
+   `githubdeploy-shoc-backend-tf-poc` and
+   `shoc-backend-tf-poc-runtime-boundary` when it creates the POC runtime
+   role; do not attach the generic account execution boundary to either role.
+3. Prove both tf-poc HCP assumptions and the retained POC import rehearsal
+   before touching the live-role ownership boundary.
+4. In a separately approved administrator/CDK migration, tag the existing HCP
+   apply roles first:
+   `hcptf-shoc-backend-dev` gets
+   `HcpTerraformWorkspace=shoc-backend-dev`, and
+   `hcptf-shoc-backend-staging` gets
+   `HcpTerraformWorkspace=shoc-backend-staging`. Next attach
+   `shoc-backend-dev-deploy-boundary` and
+   `shoc-backend-staging-deploy-boundary` to the exact `githubdeploy-*` roles,
+   and attach `shoc-backend-dev-runtime-boundary` /
+   `shoc-backend-staging-runtime-boundary` to the exact runtime roles. Verify
+   the boundary ceilings before adding the matching manager tag to either
+   target `githubdeploy-*` role. The POC CDK
+   creates its deploy role with `HcpTerraformWorkspace=shoc-backend-tf-poc`;
+   the substrate-created POC apply role already carries the same principal
+   tag. Verify each effective deployment action before continuing. HCP remains
+   blocked while a target tag is missing/different or the target lacks its
+   exact dedicated boundary, so a partial migration cannot authorize policy
+   writes. Complete both runtime/deploy boundary attachments before workload
+   imports. HCP apply roles deliberately have no
+   `iam:PutRolePermissionsBoundary` or boundary-policy mutation permissions.
+5. In the backend bootstrap, add Terraform `removed` blocks with
+   `destroy = false` for only the four dev/staging HCP roles and their inline
+   policies. Apply and verify Terraform state no longer owns them while all
+   four physical roles and ARNs remain unchanged.
+6. Set both contexts to `true`, synthesize with
+   `npx cdk synth terraform-substrate-external-dev`, and create a
+   CloudFormation **IMPORT** change set for the four existing
+   `AWS::IAM::Role` resources by exact role name. Do not run a normal
+   CREATE/UPDATE change set for this ownership transition. The POC gate must
+   remain true so the already-managed pair stays in the template. Import
+   records ownership; it does not update existing role properties or inline
+   policies.
+7. Run a separate, reviewed CloudFormation reconcile update after import and
+   before switching workspace credentials. Each current live role has one
+   inline policy: `shoc-backend-dev-import-plan`,
+   `shoc-backend-dev-import-apply`,
+   `shoc-backend-staging-import-plan`, or
+   `shoc-backend-staging-import-apply`. Existing descriptions and tags are
+   inventoried in the backend handoff. Reconcile those explicit differences
+   to the final baseline shape without replacing a role.
+8. After reconcile and HCP assumption proof, keep both context values committed
+   as `true`. Every subsequent normal deployment must synthesize all six
+   roles. Never return either gate to false as a rollback mechanism; Retain
+   protects the physical role but removing it from the stack abandons
+   CloudFormation ownership.
+9. Only after all imports/reconciliation complete and both context defaults
+   are permanently `true`, add `terraform-substrate-external-dev` back to the
+   external-dev workflow stack selector. Until then all substrate operations
+   are deliberate manual deploy/import actions.
+10. Retire the backend bootstrap only after the POC pair and all four imported
+   live roles are proven under this stack. Role deletion/recreation is never a
+   migration step.
+
+The external-dev apply roles intentionally omit role create/delete,
+managed-policy attach/detach, trust or boundary mutation, `iam:PassRole`, and
+secret-value APIs. IAM writes are limited to exact-role inline-policy and
+ordinary tag updates plus exact-profile tags; role descriptions remain stable
+and HCP receives no `UpdateRole` or `UpdateRoleDescription`. The SCP permits
+only the three enumerated HCP apply roles to mutate a `githubdeploy-*` role
+whose locked `HcpTerraformWorkspace` resource tag equals the caller's immutable
+principal tag. Adding or changing that manager tag remains administrator/CDK
+only. POC DNS and certificate access is tag/name constrained because their
+physical IDs are allocated by the temporary retained CDK stack before
+Terraform imports them. Dev and staging DNS writes are pinned to their existing
+hosted-zone IDs and API record names.
+
+Current compact policy-document sizes are 1,387 / 1,873 / 1,844 characters for
+the POC/dev/staging deploy boundaries and 1,176 / 1,779 / 1,656 for their
+runtime boundaries, each below IAM's 6,144-character managed-policy limit. The
+external-dev IAM guardrail SCP is 4,922 compact characters against its
+5,120-character Organizations limit; keep size assertions in every change.
+
 **HCP Terraform layout (org-level setup, console):** one org `seahaven`
 (free tier: 500 managed resources, 1 concurrent run); one HCP **project per
 AWS account** (`seahaven-prod`, `seahaven-dev`); one **workspace per stack**

bin/app.ts

@@ -31,6 +31,13 @@ const PROD_VPC_IDS = [
 
 const app = new cdk.App();
 
+const contextBoolean = (key: string): boolean => {
+  const value = app.node.tryGetContext(key);
+  if (value === true || value === "true") return true;
+  if (value === false || value === "false" || value === undefined) return false;
+  throw new Error(`${key} must be true or false`);
+};
+
 new AccountBaselineStack(app, "account-baseline", {
   stackName: "seahaven-account-baseline",
   env: { account: ACCOUNT, region: "us-east-1" },
@@ -190,8 +197,10 @@ const deploySubstrateDev = new DeploySubstrateStack(app, "deploy-substrate-dev",
 // OIDC provider + the shared boundary-gated guardrail policy
 // (seahaven-hcptf-iam-management) that per-workspace apply roles attach.
 // Per-workspace hcptf-* roles are appended to the template at each stack's
-// migration time, never here. prod/dev ONLY — mgmt stays SAM (Terraform POC
+// migration time, never here. prod/dev/external-dev ONLY — mgmt stays SAM
-// decision 2026-07-30; the mgmt POC substrate was rolled back the same day).
+// (Terraform POC decision 2026-07-30; the mgmt POC substrate was rolled back
+// the same day). External-dev references its existing provider and uses
+// workload-specific inline policies instead of the shared IAM manager.
 // The guardrail policy names the seahaven-lambda-execution-boundary ARN only
 // inside Condition strings, so CFN infers no creation edge — the explicit
 // dependency below guarantees the deploy-substrate stack (which owns the
@@ -227,6 +236,21 @@ const terraformSubstrateDev = new TerraformSubstrateStack(
 );
 terraformSubstrateDev.addStackDependency(deploySubstrateDev);
 
+// External-dev already has app.terraform.io federation. Both role gates start
+// false in cdk.json: POC is enabled by a normal update; dev/staging only by
+// CloudFormation import after Terraform relinquishes those four live roles.
+const terraformSubstrateExternalDev = new TerraformSubstrateStack(
+  app,
+  "terraform-substrate-external-dev",
+  {
+    stackName: "seahaven-terraform-substrate",
+    env: { account: EXTERNAL_DEV_ACCOUNT, region: "us-east-1" },
+    createOidcProvider: false,
+    enableShocBackendPocRoles: contextBoolean("enableShocBackendPocRoles"),
+    enableShocBackendLiveRoles: contextBoolean("enableShocBackendLiveRoles"),
+  },
+);
+
 // ── Shared DynamoDB CMK (INFRA-95 / M-3) ─────────────────────────────────────
 // Dedicated, standalone stack so the customer-managed key for sensitive
 // finance/PII DynamoDB tables is an independent shared dependency for the owning

cdk.json

@@ -17,6 +17,8 @@
   "context": {
     "@aws-cdk/aws-lambda:recognizeLayerVersion": true,
     "@aws-cdk/core:checkSecretUsage": true,
-    "@aws-cdk/core:target-partitions": ["aws"]
+    "@aws-cdk/core:target-partitions": ["aws"],
+    "enableShocBackendPocRoles": false,
+    "enableShocBackendLiveRoles": false
   }
 }

lib/scp/external-dev-iam-guardrails.json

@@ -12,11 +12,18 @@
       }
     },
     {
-      "Sid": "ProtectBoundaryPolicyFromEdits",
+      "Sid": "ProtectExecutionBoundary",
       "Effect": "Deny",
       "Action": ["iam:CreatePolicyVersion", "iam:SetDefaultPolicyVersion", "iam:DeletePolicy", "iam:DeletePolicyVersion"],
       "Resource": "arn:aws:iam::396287094661:policy/external-dev-execution-boundary"
     },
+    {
+      "Sid": "ProtectShocBoundaries",
+      "Effect": "Deny",
+      "Action": ["iam:CreatePolicyVersion", "iam:SetDefaultPolicyVersion", "iam:DeletePolicy", "iam:DeletePolicyVersion"],
+      "Resource": "arn:aws:iam::396287094661:policy/shoc-backend-*-boundary",
+      "Condition": { "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole", "arn:aws:iam::396287094661:role/cdk-hnb659fds-*"] } }
+    },
     {
       "Sid": "DenyAlteringPermissionsBoundaries",
       "Effect": "Deny",
@@ -32,18 +39,61 @@
       "Condition": { "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole"] } }
     },
     {
-      "Sid": "ProtectPrivilegedRoles",
+      "Sid": "ProtectNonGithubPrivilegedRoles",
       "Effect": "Deny",
       "Action": ["iam:UpdateAssumeRolePolicy", "iam:AttachRolePolicy", "iam:DetachRolePolicy", "iam:PutRolePolicy", "iam:DeleteRolePolicy", "iam:DeleteRole", "iam:UpdateRole", "iam:TagRole", "iam:UntagRole"],
       "Resource": [
         "arn:aws:iam::396287094661:role/OrganizationAccountAccessRole",
         "arn:aws:iam::396287094661:role/cdk-hnb659fds-*",
-        "arn:aws:iam::396287094661:role/githubdeploy-*",
         "arn:aws:iam::396287094661:role/seahaven-extdev-config-recorder-role",
         "arn:aws:iam::396287094661:role/seahaven-extdev-config-custom-resource-role",
         "arn:aws:iam::396287094661:role/aws-service-role/*"
       ],
       "Condition": { "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole", "arn:aws:iam::396287094661:role/cdk-hnb659fds-*"] } }
+    },
+    {
+      "Sid": "ProtectDeploymentPrincipalLifecycle",
+      "Effect": "Deny",
+      "Action": ["iam:UpdateAssumeRolePolicy", "iam:AttachRolePolicy", "iam:DetachRolePolicy", "iam:DeleteRolePolicy", "iam:DeleteRole", "iam:UpdateRole", "iam:UpdateRoleDescription", "iam:PutRolePermissionsBoundary", "iam:DeleteRolePermissionsBoundary"],
+      "Resource": ["arn:aws:iam::396287094661:role/githubdeploy-*", "arn:aws:iam::396287094661:role/hcptf-*"],
+      "Condition": { "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole", "arn:aws:iam::396287094661:role/cdk-hnb659fds-*"] } }
+    },
+    {
+      "Sid": "ProtectGithubRoleMetadata",
+      "Effect": "Deny",
+      "Action": ["iam:PutRolePolicy", "iam:TagRole", "iam:UntagRole"],
+      "Resource": "arn:aws:iam::396287094661:role/githubdeploy-*",
+      "Condition": { "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole", "arn:aws:iam::396287094661:role/cdk-hnb659fds-*", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-tf-poc", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-dev", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-staging"] } }
+    },
+    {
+      "Sid": "DenyUnmanagedGithubRole",
+      "Effect": "Deny",
+      "Action": ["iam:PutRolePolicy", "iam:TagRole", "iam:UntagRole"],
+      "Resource": "arn:aws:iam::396287094661:role/githubdeploy-*",
+      "Condition": {
+        "ArnLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/hcptf-shoc-backend-tf-poc", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-dev", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-staging"] },
+        "Null": { "aws:ResourceTag/HcpTerraformWorkspace": "true" }
+      }
+    },
+    {
+      "Sid": "EnforceGithubRoleManagerTag",
+      "Effect": "Deny",
+      "Action": ["iam:PutRolePolicy", "iam:TagRole", "iam:UntagRole"],
+      "Resource": "arn:aws:iam::396287094661:role/githubdeploy-*",
+      "Condition": {
+        "ArnLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/hcptf-shoc-backend-tf-poc", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-dev", "arn:aws:iam::396287094661:role/hcptf-shoc-backend-staging"] },
+        "StringNotEquals": { "aws:ResourceTag/HcpTerraformWorkspace": "${aws:PrincipalTag/HcpTerraformWorkspace}" }
+      }
+    },
+    {
+      "Sid": "LockHcpTerraformWorkspaceTag",
+      "Effect": "Deny",
+      "Action": ["iam:TagRole", "iam:UntagRole"],
+      "Resource": "arn:aws:iam::396287094661:role/*",
+      "Condition": {
+        "ArnNotLike": { "aws:PrincipalArn": ["arn:aws:iam::396287094661:role/OrganizationAccountAccessRole", "arn:aws:iam::396287094661:role/cdk-hnb659fds-*"] },
+        "ForAnyValue:StringEquals": { "aws:TagKeys": "HcpTerraformWorkspace" }
+      }
     }
   ]
 }

lib/terraform-substrate-stack.ts

@@ -17,14 +17,26 @@ export interface TerraformSubstrateStackProps extends cdk.StackProps {
    * remove the orphaned provider or redeploy with this false.
    */
   createOidcProvider?: boolean;
+
+  /**
+   * Enable the two SHOC backend tf-poc HCP roles. Defaults false so the
+   * external-dev base-stack create is role-free.
+   */
+  enableShocBackendPocRoles?: boolean;
+
+  /**
+   * Enable the four existing SHOC backend dev/staging HCP roles. Defaults
+   * false because these names must enter the stack through CloudFormation
+   * resource import, never a normal create/update.
+   */
+  enableShocBackendLiveRoles?: boolean;
 }
 
 /**
- * Per-account HCP Terraform deploy substrate: the shared account-level
+ * Per-account HCP Terraform deploy substrate: the conditional
- * resources every Terraform workspace pipeline needs -
+ * app.terraform.io OIDC provider, the prod/dev shared boundary-gated IAM
- *  - app.terraform.io OIDC identity provider (conditional, see props), and
+ * manager, and reviewed per-workspace role pairs. External-dev conditions out
- *  - `seahaven-hcptf-iam-management`, the shared boundary-gated IAM
+ * the shared manager and uses exact inline policies for its SHOC import roles.
- *    guardrail policy every per-workspace APPLY role attaches.
  *
  * Deliberately NOT here: per-workspace hcptf-<stack> / hcptf-<stack>-plan
  * roles. Those are appended to the template at each stack's migration time
@@ -58,6 +70,10 @@ export class TerraformSubstrateStack extends cdk.Stack {
       ),
       parameters: {
         CreateOIDCProvider: props?.createOidcProvider === false ? "false" : "true",
+        EnableShocBackendPocRoles:
+          props?.enableShocBackendPocRoles === true ? "true" : "false",
+        EnableShocBackendLiveRoles:
+          props?.enableShocBackendLiveRoles === true ? "true" : "false",
       },
     });