Merge milestone/v1.16-nova-simplification — v1.16 complete (Nova Simplification: 20-phase NFR sweep + final; tag v1.15.26)

core/lambda/contract_ingestor.py

@@ -30,10 +30,53 @@ PLATFORM_REPO = os.environ.get("PLATFORM_REPO", "nova/acdl")
 # to a Gitea API root (e.g. https://git.cloudinit.dev/api/v1) for Gitea.
 GITHUB_API_BASE = os.environ.get("GITHUB_API_BASE", "https://api.github.com")
 
+# P11 (REQ-175): consistent cap for error/stackTrace fields (was 10k vs 2k).
+MAX_ERROR_FIELD_CHARS = 10000
+# P11 (REQ-175): max contract blob size before the DynamoDB write (256 KB).
+MAX_CONTRACT_BYTES = 256 * 1024
+
 _dynamodb = None
 _secrets_client = None
 
 
+def _discover_environments():
+    """P10 (REQ-174): derive the valid environment names from
+    core/environments/*.json (the directory is the single source of truth,
+    not a hardcoded set). Falls back to {'dev','qa','prod','dr'} if the
+    directory is not readable (e.g. packaged Lambda without the dir).
+    """
+    env_dir = os.path.join(os.path.dirname(os.path.dirname(os.path.dirname(
+        os.path.abspath(__file__)))), "core", "environments")
+    try:
+        names = {f[:-5] for f in os.listdir(env_dir) if f.endswith(".json")}
+        return names or {"dev", "qa", "prod", "dr"}
+    except OSError:
+        return {"dev", "qa", "prod", "dr"}
+
+
+def _validate_contract_schema(contract):
+    """P11 (REQ-175): validate the contract blob against
+    schemas/contract.schema.json before the DynamoDB write. Raises
+    ValueError on invalid. Falls back to a no-op if the schema or
+    jsonschema is unavailable (e.g. packaged Lambda without the schema).
+    """
+    try:
+        import json as _json
+        import jsonschema
+        schema_path = os.path.join(os.path.dirname(os.path.dirname(
+            os.path.dirname(os.path.abspath(__file__)))),
+            "schemas", "contract.schema.json")
+        with open(schema_path) as f:
+            schema = _json.load(f)
+        jsonschema.validate(instance=contract, schema=schema)
+    except (OSError, ImportError):
+        # Schema or jsonschema unavailable — no-op (the contract is
+        # validated upstream by run_platform.sh in the normal path).
+        pass
+    except jsonschema.ValidationError as e:
+        raise ValueError(f"contract schema validation failed: {e.message}")
+
+
 def _get_dynamodb():
     global _dynamodb
     if _dynamodb is None:
@@ -94,6 +137,25 @@ def _submit_contract(payload):
     contract_id = payload["contractId"]
     contract = payload["contract"]
     environment = payload["environment"]
+
+    # P11 (REQ-175): size-cap the contract blob before the DynamoDB write
+    # (unbounded payload → write amplification). 256 KB matches DynamoDB
+    # item limit headroom; reject oversized with a clear error.
+    import json as _json
+    contract_json = _json.dumps(contract).encode()
+    if len(contract_json) > MAX_CONTRACT_BYTES:
+        raise ValueError(
+            f"contract payload too large: {len(contract_json)} bytes "
+            f"(max {MAX_CONTRACT_BYTES} bytes / 256 KB)"
+        )
+
+    # P11 (REQ-175): schema-validate the contract blob against
+    # schemas/contract.schema.json before the write. Reject invalid with 400.
+    # The local Lambda stub (NOVA_LAMBDA_LOCAL_BYPASS) skips schema validation
+    # — it tests the invoke path, not real contract submission.
+    if not os.environ.get("NOVA_LAMBDA_LOCAL_BYPASS"):
+        _validate_contract_schema(contract)
+
     submitted_at = _iso8601_now()
     table = _get_dynamodb().Table(TABLE_NAME)
     item = {
@@ -131,7 +193,7 @@ def _report_error(payload):
     contract_id = payload["contractId"]
     error = payload.get("error", "unknown error")
     run_url = payload.get("runUrl", "")
-    stack_trace = payload.get("stackTrace", "")[:2000]  # truncate
+    stack_trace = payload.get("stackTrace", "")[:MAX_ERROR_FIELD_CHARS]  # P11: aligned cap
 
     # Get the GitHub token from Secrets Manager
     secrets = _get_secrets_client()
@@ -142,7 +204,7 @@ def _report_error(payload):
         raise RuntimeError(f"failed to read GitHub token from Secrets Manager: {e}")
 
     owner, repo = PLATFORM_REPO.split("/")
-    title = f"[ACDL-ALERT] Deploy failure: {consumer_repo} / {contract_id}"
+    title = f"[NOVA-ALERT] Deploy failure: {consumer_repo} / {contract_id}"
 
     # Check for an existing open issue with the same title (idempotency)
     # URL-encode the contract_id to prevent search-query injection (P1-1).
@@ -188,7 +250,7 @@ def _report_error(payload):
 {stack_trace}
 ```
 
-_This issue was auto-created by the ACDL platform Lambda (D-055). The consumer's onboarding-granted Lambda-invoke permission is the only grant needed._
+_This issue was auto-created by the Nova platform Lambda (D-055). The consumer's onboarding-granted Lambda-invoke permission is the only grant needed._
 """
 
     if existing:
@@ -236,20 +298,33 @@ def _validate_caller_identity(event, payload):
     in the payload matches the principal's ARN-derived source identity, preventing
     one consumer from impersonating another.
 
-    If the identity is not available (e.g. local testing or non-IAM auth), the
-    check is skipped (the ABAC policy at the IAM layer enforces the scope).
+    P10 (REQ-174): if the IAM identity is absent (no callerArn), the function
+    FAILS CLOSED (raises ValueError) rather than silently passing. The ABAC
+    policy at the IAM layer is the primary enforcement; this is defense-in-
+    depth so a misconfigured Function URL (no IAM auth) does not allow
+    unauthenticated contract submission. Local testing must set a test ARN
+    via the event requestContext or the LOCAL_LAMBDA_STUB env bypass.
 
     v1.14 (REQ-144): also validates contractId format, environment enum, and
-    error length. The ABAC reliance is documented here: the Function URL IAM
-    identity does not expose principal tags in the event, so full enforcement
-    of consumerRepo ownership is at the IAM layer (ABAC via
-     aws:PrincipalTag/nova:owner). This function validates format only, not
-    ownership.
+    error length. P10 (REQ-174): the environment enum is derived from the
+    core/environments/ directory (not hardcoded), so a new env JSON is the
+    single source of truth. The ABAC reliance is documented here: the
+    Function URL IAM identity does not expose principal tags in the event,
+    so full enforcement of consumerRepo ownership is at the IAM layer (ABAC
+    via aws:PrincipalTag/nova:owner). This function validates format only,
+    not ownership.
     """
     identity = event.get("requestContext", {}).get("identity", {})
     caller_arn = identity.get("userArn", "")
     if not caller_arn:
-        pass  # no identity available — rely on IAM ABAC enforcement
+        # P10 (REQ-174): fail closed. A local-test bypass is allowed via
+        # the NOVA_LAMBDA_LOCAL_BYPASS env var (set by the LocalLambdaStub).
+        import os as _os
+        if not _os.environ.get("NOVA_LAMBDA_LOCAL_BYPASS"):
+            raise ValueError(
+                "missing IAM caller identity (requestContext.identity.userArn) — "
+                "the Function URL must use IAM auth; refusing unauthenticated submission"
+            )
     payload_repo = payload.get("consumerRepo", "")
     if payload_repo:
         # consumerRepo must be org/repo format, <=128 chars
@@ -263,17 +338,18 @@ def _validate_caller_identity(event, payload):
         if not re.match(r'^[a-zA-Z0-9][a-zA-Z0-9_-]{0,63}$', contract_id):
             raise ValueError(f"invalid contractId format: {contract_id!r} (alphanumeric, hyphen, underscore; max 64 chars)")
 
-    # v1.14 (REQ-144): environment enum validation
+    # P10 (REQ-174): environment enum derived from core/environments/ (not
+    # hardcoded) — the directory is the single source of truth.
     environment = payload.get("environment", "")
     if environment:
-        valid_envs = {"dev", "qa", "prod", "dr"}
+        valid_envs = _discover_environments()
         if environment not in valid_envs:
-            raise ValueError(f"invalid environment: {environment!r} (must be one of {valid_envs})")
+            raise ValueError(f"invalid environment: {environment!r} (must be one of {sorted(valid_envs)})")
 
     # v1.14 (REQ-144): error length cap (for report_error action)
     error_msg = payload.get("error", "")
-    if error_msg and len(str(error_msg)) > 10000:
-        payload["error"] = str(error_msg)[:10000]
+    if error_msg and len(str(error_msg)) > MAX_ERROR_FIELD_CHARS:
+        payload["error"] = str(error_msg)[:MAX_ERROR_FIELD_CHARS]
 
 
 def _validate_change_request(payload):
@@ -322,6 +398,65 @@ def _validate_change_request(payload):
     }
 
 
+def _onboard_consumer(payload):
+    """P18 (REQ-182): accept a self-service onboarding request.
+
+    Validates the payload against schemas/onboarding.schema.json, then
+    writes a 'pending' row to nova-contracts (D-119). No AWS resources
+    are created by this action (D-113); the cross-account role + ABAC
+    tag grant is offline-proven Terraform (P20/REQ-184).
+    """
+    import jsonschema
+    schema_path = os.path.join(os.path.dirname(os.path.dirname(
+        os.path.dirname(os.path.abspath(__file__)))),
+        "schemas", "onboarding.schema.json")
+    try:
+        with open(schema_path) as f:
+            schema = json.load(f)
+        # Strip the Lambda dispatch envelope (action) before validating
+        # against the onboarding schema (the schema is about the request,
+        # not the Lambda wrapper).
+        onboarding_payload = {k: v for k, v in payload.items() if k != "action"}
+        jsonschema.validate(instance=onboarding_payload, schema=schema)
+    except OSError:
+        raise ValueError("onboarding schema unavailable")
+    except jsonschema.ValidationError as e:
+        raise ValueError(f"onboarding payload invalid: {e.message}")
+
+    consumer_repo = payload["consumerRepo"]
+    requested_env = payload["requestedEnvironment"]
+    owner_id = payload["ownerId"]
+    billing_tag = payload["billingTag"]
+    submitted_at = _iso8601_now()
+
+    # Write a pending CMDB row (PK consumerRepo, SK onboarding#env#timestamp).
+    table = _get_dynamodb().Table(TABLE_NAME)
+    item = {
+        "consumerRepo": consumer_repo,
+        "contractId#submittedAt": f"onboarding#{requested_env}#{submitted_at}",
+        "contractId": f"onboarding-{requested_env}",
+        "environment": requested_env,
+        "status": "pending",
+        "ownerId": owner_id,
+        "billingTag": billing_tag,
+        "notes": payload.get("notes", ""),
+        "submittedAt": submitted_at,
+    }
+    table.put_item(TableName=TABLE_NAME, Item=item)
+    return {
+        "status": "pending",
+        "consumerRepo": consumer_repo,
+        "requestedEnvironment": requested_env,
+        "action": "onboard_consumer",
+        "submittedAt": submitted_at,
+        "message": (
+            "Onboarding request received. The platform team will provision "
+            "the environment binding + cross-account role. Track the status "
+            "via the nova-contracts table (status=pending → granted)."
+        ),
+    }
+
+
 def lambda_handler(event, context):
     """AWS Lambda handler entry point.
 
@@ -350,6 +485,8 @@ def lambda_handler(event, context):
             result = _report_error(payload)
         elif action == "validate_change_request":
             result = _validate_change_request(payload)
+        elif action == "onboard_consumer":
+            result = _onboard_consumer(payload)
         else:
             return {
                 "statusCode": 400,
@@ -357,6 +494,9 @@ def lambda_handler(event, context):
             }
         return {"statusCode": 200, "body": json.dumps(result)}
     except ValueError as e:
+        # P10 (REQ-174): identity failures are 401, field validation is 400.
+        if "missing IAM caller identity" in str(e):
+            return {"statusCode": 401, "body": json.dumps({"error": str(e)})}
         return {"statusCode": 400, "body": json.dumps({"error": str(e)})}
     except Exception as e:  # pragma: no cover - defensive top-level guard
         return {"statusCode": 500, "body": json.dumps({"error": str(e)})}