Configure the assessment platform as the LTI platform and the proctoring service as the LTI tool; register the proctoring tool using LTI 1.3 dynamic registration or manual key exchange
When the learner starts an exam, the assessment platform sends a proctoring start request message to the proctoring tool as an LTI launch; the tool performs identity verification and environment checks then returns a start proctoring response message
The proctoring tool sends a tool-originating end assessment message back to the platform when the proctoring session begins, signaling the platform to allow the learner to access the assessment content
During the exam the proctoring tool monitors the candidate; upon exam completion or irregularity detection, the tool sends the appropriate end proctoring message to the platform
The assessment platform uses the AGS service to receive the final proctoring result if the proctoring tool posts a score or status back via AGS after human review
Known gotchas
The proctoring specification defines a bi-directional message flow; an assessment platform that only implements the outbound start request and does not handle inbound tool-originating messages will stall the exam session waiting for a signal that never comes
The candidate must remain in the proctoring tool's context during the exam; navigating away or closing the proctor window before the end-assessment signal is sent may leave the exam in an indeterminate state in the platform
The specification requires specific JWT claims for each message type; generic LTI 1.3 launch handling that ignores the message_type claim will conflate proctoring messages with normal resource link launches
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp