Harden external file connector boundaries

This commit is contained in:
2026-07-21 12:10:23 +02:00
parent 3bc1d3489e
commit f2dfb6c90e
18 changed files with 1167 additions and 64 deletions

View File

@@ -0,0 +1,199 @@
from __future__ import annotations
import os
import re
from collections.abc import Mapping
from pathlib import Path
from typing import Any
_SECRET_ENV_ALLOWLIST = "GOVOPLAN_CONNECTOR_SECRET_ENV_ALLOWLIST" # noqa: S105 # nosec B105 - configuration key.
_CA_BUNDLE_ALLOWLIST = "GOVOPLAN_CONNECTOR_CA_BUNDLE_ALLOWLIST"
_ENV_NAME = re.compile(r"[A-Za-z_][A-Za-z0-9_]*\Z")
_SECRET_ENV_METADATA_KEYS = frozenset(
{
"access_key_id_env",
"secret_access_key_env",
"session_token_env",
}
)
_DEVELOPMENT_ENVIRONMENTS = frozenset({"dev", "development", "local", "test", "testing"})
class ConnectorDeploymentConfigurationError(ValueError):
"""Raised when connector data crosses a deployment-owned trust boundary."""
def connector_secret_env_value(name: str, *, source_kind: str) -> str:
clean_name = _validate_secret_env_reference(name, source_kind=source_kind)
value = os.environ.get(clean_name)
if value is None or value == "":
raise ConnectorDeploymentConfigurationError("Connector credential environment variable is not configured")
return value
def connector_secret_env_available(name: str | None, *, source_kind: str) -> bool:
if not name:
return False
try:
clean_name = _validate_secret_env_reference(name, source_kind=source_kind)
except ConnectorDeploymentConfigurationError:
return False
return bool(os.environ.get(clean_name))
def validate_deployment_connector_references(
*,
source_kind: str,
password_env: str | None = None,
token_env: str | None = None,
metadata: Mapping[str, Any] | None = None,
) -> None:
"""Validate references in deployment-owned connector configuration."""
for name in (password_env, token_env):
if name:
_validate_secret_env_reference(name, source_kind=source_kind)
for key in _SECRET_ENV_METADATA_KEYS:
name = _clean((metadata or {}).get(key))
if name:
_validate_secret_env_reference(name, source_kind=source_kind)
validate_connector_tls_metadata(metadata)
def reject_api_controlled_deployment_references(
*,
password_env: str | None = None,
token_env: str | None = None,
metadata: Mapping[str, Any] | None = None,
) -> None:
"""Reject process-secret selectors controlled through tenant-facing APIs."""
if _clean(password_env) or _clean(token_env):
raise ConnectorDeploymentConfigurationError(
"Environment-backed credentials may only be declared in deployment-owned connector configuration"
)
for key, value in (metadata or {}).items():
if _clean(value) and (str(key).strip().casefold() in _SECRET_ENV_METADATA_KEYS or str(key).strip().casefold().endswith("_env")):
raise ConnectorDeploymentConfigurationError(
"Environment-backed credentials may only be declared in deployment-owned connector configuration"
)
validate_connector_tls_metadata(metadata)
def connector_ca_bundle_path(value: str) -> str:
clean_value = _clean(value)
if not clean_value:
raise ConnectorDeploymentConfigurationError("Connector CA bundle path is empty")
candidate = Path(clean_value)
if not candidate.is_absolute():
raise ConnectorDeploymentConfigurationError("Connector CA bundle paths must be absolute")
try:
resolved = candidate.resolve(strict=True)
except OSError as exc:
raise ConnectorDeploymentConfigurationError("Connector CA bundle path does not exist") from exc
allowed = _allowed_ca_bundle_paths()
if resolved not in allowed:
raise ConnectorDeploymentConfigurationError(
f"Connector CA bundle path is not listed in {_CA_BUNDLE_ALLOWLIST}"
)
if not resolved.is_file():
raise ConnectorDeploymentConfigurationError("Connector CA bundle path must be a regular file")
return str(resolved)
def validate_connector_tls_metadata(metadata: Mapping[str, Any] | None) -> None:
values = metadata or {}
ca_bundle = _clean(values.get("ca_bundle"))
if ca_bundle:
connector_ca_bundle_path(ca_bundle)
for key in ("verify_tls", "tls_verify"):
if key in values and not _as_bool(values.get(key), default=True) and not _development_runtime():
raise ConnectorDeploymentConfigurationError(
"Connector TLS certificate verification may only be disabled in dev/test environments"
)
def connector_effective_endpoint_url(
*,
provider: str | None,
endpoint_url: str | None,
metadata: Mapping[str, Any] | None,
) -> str | None:
"""Return the endpoint that connector I/O will actually use."""
clean_provider = (provider or "").strip().casefold()
values = metadata or {}
webdav_url = _clean(values.get("webdav_endpoint_url"))
browse_protocol = (_clean(values.get("browse_protocol")) or "").casefold()
if webdav_url and (clean_provider in {"seafile", "webdav", "nextcloud"} or browse_protocol == "webdav"):
return webdav_url
return _clean(endpoint_url)
def _validate_secret_env_reference(name: str, *, source_kind: str) -> str:
if source_kind.strip().casefold() != "settings":
raise ConnectorDeploymentConfigurationError(
"Environment-backed credentials may only be used by deployment-owned connector configuration"
)
clean_name = name.strip()
if not _ENV_NAME.fullmatch(clean_name):
raise ConnectorDeploymentConfigurationError("Connector credential environment variable name is invalid")
if clean_name not in _allowed_secret_env_names():
raise ConnectorDeploymentConfigurationError(
f"Connector credential environment variable is not listed in {_SECRET_ENV_ALLOWLIST}"
)
return clean_name
def _allowed_secret_env_names() -> frozenset[str]:
raw = os.environ.get(_SECRET_ENV_ALLOWLIST, "")
names = frozenset(item.strip() for item in raw.split(",") if item.strip())
invalid = sorted(name for name in names if not _ENV_NAME.fullmatch(name))
if invalid:
raise ConnectorDeploymentConfigurationError(f"{_SECRET_ENV_ALLOWLIST} contains an invalid variable name")
return names
def _allowed_ca_bundle_paths() -> frozenset[Path]:
raw = os.environ.get(_CA_BUNDLE_ALLOWLIST, "")
paths: set[Path] = set()
for item in (part.strip() for part in raw.split(",")):
if not item:
continue
path = Path(item)
if not path.is_absolute():
raise ConnectorDeploymentConfigurationError(f"{_CA_BUNDLE_ALLOWLIST} requires absolute paths")
try:
paths.add(path.resolve(strict=True))
except OSError as exc:
raise ConnectorDeploymentConfigurationError(f"{_CA_BUNDLE_ALLOWLIST} contains a missing path") from exc
return frozenset(paths)
def _development_runtime() -> bool:
app_env = os.environ.get("APP_ENV", "dev").strip().casefold()
install_profile = os.environ.get("GOVOPLAN_INSTALL_PROFILE", "").strip().casefold()
return app_env in _DEVELOPMENT_ENVIRONMENTS and (
not install_profile or install_profile in _DEVELOPMENT_ENVIRONMENTS
)
def _as_bool(value: object, *, default: bool) -> bool:
if value is None:
return default
if isinstance(value, bool):
return value
clean = str(value).strip().casefold()
if clean in {"1", "true", "yes", "on"}:
return True
if clean in {"0", "false", "no", "off"}:
return False
raise ConnectorDeploymentConfigurationError("Connector TLS verification setting must be true or false")
def _clean(value: object) -> str | None:
if value is None:
return None
clean = str(value).strip()
return clean or None