Command surface for the whole system.
Routines stay first, but the root dashboard gives a wider operational map: attention, runs, entry points and the next layers we will plug in.
Core modules
Attention
Top routines
Routine table, first class.
This is the operational table the team will use daily. It reads the same signals as the CLI and exposes safe dry-run actions.
Routines overview
Operational ledger
Delegate runs, visible.
Inspect which executor was chosen, whether the run passed, and open the full
artifact without dropping to terminal or manual cat.
Recent activity signal
Delegate runs overview
Operational posture and alarms.
This page isolates the PRANA health surface: routine status, active attention and schedule drift.
Health breakdown
Attention queue
Ready to run
Learnings, drift and sync state.
This page surfaces the memory snapshot used by the operator loop, including open todos and external signals.
Open items
Cross-project signals
Memory shortcuts
Stories, backlog and execution queue.
This page turns the story ledger into a working queue for the PRANA dashboard: open work, review and fresh items.
Review queue
Active stories
Latest ledger entries
Autonomous session under supervision.
This view observes the latest `aios-master` session: state, next action, handoffs and compact context.
Current route
Master directive
Policy and budget
Request to session trace
Recent executions
Delegation trail
Jobs and trust trail
Read-only artifacts
Recent session events
Resume prompt
Request understanding before execution.
This view observes the latest Master request: conversation, understanding, story draft, approval and lineage.
Request to session trace
What was approved
Scope and evidence
Latest request
Recent messages
Master understanding
Execution candidate
Recent Master requests
Heterogeneous router efficiency.
Tracks eval/utility offloads to Azure, semantic cache hits, and estimated savings vs running everything on Claude.