For AI agents: a documentation index is available at the root level at /llms.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
Version 8.7.0
- Use
undici as the HTTP client instead of Phin, which is no longer maintained.
- Added
undici HTTP client implementation
- Added
risk_rtt field to activities in case of S2S timeout
- Support for JWT key names that contain dots (such as URLs) when extracting the user id and additional fields
- Custom block on error message configuration support
- Pass request object to response custom parameters custom function
- MCP Protection support
- Send original vid from cookie on all activities under orig_cookie_vid
- Added Hard Block Support
Version 8.6.0
- Changed response headers being set as arrays instead of strings
Version 8.5.0
- Added support for
custom_proxy feature (px_proxy_url configuration)
Version 8.4.1
- Added support for the
custom_sensitive_request feature, which enables using a custom function to identify sensitive routes
- Added
is_sensitive_route field in activities
Version 8.3.0
- Added support for running the Enforcer as a standalone Node.js process without requiring a code bundle.
- Added
remote_config_id field to risk and async activities
- Added
remote_config_id and remote_config_version fields to telemetry activity
DefaultGraphQLParser modified to extract operationName and variables even when the query is not present
- Specifying token version (v2 or v3) with
px_token_version always for consistency in telemetry activities
- Fixed issue where telemetry would fail for array of cookie strings
- Fixed issue where
UrlImpl constructor would throw an error for @ character in the path or search params
- CD process fix
Version 8.2.1
- Decode the request URL, send it on the enforcer activities URL field, and use it for all enforcer processing.
Version 8.2.0
- Added
px_token_version configuration to support both v2 and v3 (default is v3)
- GraphQL query keyword extraction via string/regex (
px_graphql_keywords) and custom function (px_extract_graphql_keywords)
- Support for cookie secret rotation
- Configuration
px_sensitive_graphql_operation_names expanded to include regular expressions and applies to extracted GraphQL keywords as well
- Telemetry activity includes redacted sensitive configuration fields
- Default value for
px_bypass_monitor_header changed from an empty string to “x-px-block”
- Modify telemetry activity to include all types of config
- Updated the captcha template
- Using raw URL instead of parsed URL in block page captcha script query parameter
- Converted fields
login_successful_reporting_method, sent_through of CredentialEndpointConfiguration to optional
- Fixed issue where unvalidated
_pxvid value was added to the captcha page
- Fixed issue where regular expressions occasionally failed on calls to
test() due to global flag
Version 8.1.0
- Added base64-encoded request HTTP method to captcha script query parameters on block pages
- JSON parsing issue with generated package.json for CommonJS library build fixed
- Issue with mix-ups in header-based logs fixed
- Issue where block activities may contain
http_status_code field fixed
Version 8.0.1
- Issue with duplicate sending of batched activities fixed
- Memory issue with header-based logger fixed
Version 8.0.0
- Refactor to base on JS Core library
- Maintains support for:
- Additional activity handler
- Advanced blocking response
- Block activity
- Block page captcha
- Block page rate limit
- Bypass monitor header
- Client IP extraction
- Cookie v3
- CORS support
- Credentials intelligence
- CSS ref
- Custom cookie header
- Custom first-party endpoints
- Custom logo
- Custom parameters
- Enforced routes
- Filter by extension
- Filter by HTTP method
- Filter by IP
- Filter by route
- Filter by user agent
- First party
- GraphQL support
- Header-based logger
- Hype sale challenge
- JS ref
- Logger
- Mobile support
- Module enable
- Module mode
- Monitored routes
- Page requested activity
- PXDE
- PXHD
- Risk API
- Sensitive headers
- Sensitive routes
- Telemetry command
- URL decode reserved characters
- User identifiers
- VID extraction