What's New in OpenEdge MCP Server 1.1
- Last Updated: August 12, 2026
- 2 minute read
- OpenEdge MCP Server
- Version 1.1
|
OpenEdge MCP Server 1.1 delivers the following enhancements: |
Cryptographic User Authentication with JWKS Validation
OpenEdge MCP Server now supports cryptographic validation of user JSON Web Tokens (JWTs) against a remote JSON Web Key Set (JWKS) endpoint, enabling OAuth2/OIDC-compliant authentication. This enhancement replaces structural-only checks with full RSA signature verification and works with identity providers like Keycloak, Auth0, Azure AD, and Okta.
When user_auth.jwks_uri is configured, user JWTs are validated for
signature verification, expiration and timing claims, issuer and audience validation
against your configured values, and algorithm verification against your explicit
allowlist, with scope extraction for fine-grained authorization. Tokens failing
validation are rejected outright, with no fallback.
Existing deployments without JWKS configuration continue in passthrough mode—no
changes required unless you opt in. Without user_auth.jwks_uri,
user_auth.mode: "jwt" performs structural validation only,
without establishing trusted claims or scopes.
For more information, see Configure user authentication.
File-Based Tool Arguments
mcpgen client command now accepts tool arguments from a file, avoiding
fragile shell quoting for large or complex JSON payloads. You can use either
--args-file or --args with the
@ prefix to pass JSON arguments from a file instead of inline
on the command line. For example, use:
|
|
mcpgen was invoked. For more information, see Client.
JSON Output Mode for CLI
The mcpgen client now provides clean JSON-only
output when using the --json flag, with all non-JSON progress
messages suppressed. You can use mcpgen --json commands in CI/CD
pipelines and automation scripts with confidence that the output contains only valid
JSON, eliminating the need for complex output filtering. This enhancement provides
reliable machine-readable output for automation pipelines without extraneous output
from progress messages or warnings, simplifying integration with JSON parsing tools
like jq.
For more information, see Client.
Request Timeout Adjustment
The default request timeout has been increased from 500ms to 30 seconds to better accommodate
real-world API workload patterns. This change prevents healthy calls to remote
OpenEdge services from failing under normal network latency. Deployments
experiencing frequent timeout errors should see improved reliability without
configuration changes. However, this changes the maximum time an unconfigured
request can occupy server resources. You can also set an explicit value using
runtime.middleware.request_timeout.timeout_ms in your
configuration. Progress recommends that you set a value based on the latency of your
downstream service and your caller's timeout requirements.
For more information, see Configure middleware settings.