LPM's AI review found this version suspicious. That disagrees with the OpenSSF/OSV advisory, which supplies the block.
The user-triggered dashboard exposes unauthenticated routes that run project scripts and modify planning files, and its listener does not restrict the host. This is a reachable dashboard security flaw, but the invoked scripts serve the advertised features and no malicious payload is evident, so it warrants a warning rather than a publish block.
Static reason
No blocking static signals were detected.
Trigger
A user starts the documented `ui` or `dashboard` command; while it runs, a reachable network client can call the routes.
Impact
A remote client that can reach the port can trigger project helper scripts and modify planning data while the dashboard is running.
Mechanism
The HTTP server listens without restricting its host, and POST routes invoke project scripts or update planning files without checking authorization.
Evidence
bin/dashboard-server.jsbin/index.jsscripts/verify-task.shscripts/sync-contracts.jsplanning/task.mdplanning/tasks/${taskId}.mdplanning/updateplanning.md
Network endpoints4
TCP port 4040 by default (dashboard HTTP server)/api/verify/api/sync-contracts/api/tasks/update-status