SPEC LIBRARY
19 typed JSON specs across 8 domains. Each spec is a contract between agent and surface — emit the right type, get the right UI.
nodus:mit-cardMost Important Task with confidence score and deadline.
MIT
Finalize Q2 product roadmap
87% confidence
Due: Tomorrow, 5 PM
nodus:action-itemTrackable task with owner, deadline and status.
Review integration test results
in progress
Platform teamDue: Friday
nodus:quick-winsRanked list of low-effort / high-impact actions.
This week's quick wins
Update deployment runbook↑ high Add caching to slow endpoint↑ high Archive stale Linear tickets nodus:progress-barInline progress indicator with semantic color states.
nodus:approval-cardHuman-in-the-loop gate with reasoning and consequences.
Approval Required
Deploy schema migration to production
✓ All 847 integration tests passing
✓ Staging ran cleanly for 72h
94% confidence
nodus:formStructured data-collection surface with validation.
nodus:morning-briefingDaily operating surface synthesized by the briefing agent.
Tuesday, March 18
Ship the A2UI spec library
📅 10 AM — Sprint review (45 min)
⚠ Staging deploy window closes at noon
nodus:meeting-prepPre-meeting intelligence: agenda, attendees, risks.
10:00 AM
Jordan Chen92
Alex Reyes78
nodus:meeting-summaryPost-meeting distillation with action items and decisions.
Sprint Review — Week 11
84%
Actions
● Update roadmap doc
● Fix flaky test suite
nodus:text-fieldSingle-line text input with label and validation.
Project name*
Enter project name...
Used in URLs — lowercase, no spaces.
nodus:selectDropdown selection with options and helper text.
Deployment region
Select region...▾
3 options
nodus:checkboxBoolean toggle — checkbox or switch variant.
Enable distributed tracing
Adds ~2% overhead but enables full span visibility.
observability
Agent Timeline
nodus:agent-timelineStep-by-step execution trace for a running agent task.
observability
Context Window
nodus:context-windowToken budget usage with breakdown by context type.
observability
Cost Dashboard
nodus:cost-dashboardToken spend and cost breakdown by agent or session.
nodus:reasoning-treeStructured argument showing evidence for a conclusion.
Should we roll out the new rate limiter to production today?
Yes — proceed with staged rollout
Load tests show 40% headroom at peak No regressions in staging (72h soak) One edge-case failure in burst tests 88% overall
epistemics
Provenance Chain
nodus:provenance-chainSource attribution chain for an agent-generated claim.
Current MRR is $48,200 with 12% MoM growth
Sources
◻stripe-revenue.csv (Mar 17)
98%governance
Escalation Banner
nodus:escalation-bannerSurfaced exception requiring immediate human attention.
critical escalation47s ago
Agent halted — action requires explicit human authorization
deploy-agent was about to execute DROP TABLE on production. Guardrail triggered.
→ On-call engineer
governance
Learning Ledger
nodus:learning-ledgerAccumulated behavioral rules from corrections and feedback.
briefing-agent rules
Never include personal names not in vault pplgraph
correction96%×14
Label AI estimates with [AI ESTIMATE] tag
guardrail100%×41
Skip watchouts section if calendar has no conflicts
feedback78%×6