← Back to Index

Morning Briefing Redesign - Implementation Summary

Completed: February 10, 2026 (20:45 UTC)
Subagent: Minnie (gradient descent optimizer)
Status: ✅ READY FOR DEPLOYMENT

What Was Delivered

1. New Morning Briefing Template (COMPLETE)

File: working/ops/morning-briefing-redesign.md (Part 1)

Structure: 10-section template aligned to Quan's loss function

Key Differences from Current Briefing:

2. Implementation Guide (COMPLETE)

File: working/ops/morning-briefing-redesign.md (Part 2)

Cron Changes Required:

New Cron Prompt File:

Data Fetching Tools:

3. News Deduplication Schema (COMPLETE)

File: tools/news-shared.json (created + initialized)

Schema Features:

Categories Defined:

Dedup Rules (6 rules specified):

  1. Identical headline → SKIP
  2. Same source URL → SKIP
  3. Semantic similarity (>85%) + same domain → SKIP
  4. Follow-up story on same event → INCLUDE with relation note
  5. Old story (>30 days) with new developments → INCLUDE with revision note
  6. Context mention (not news item) → INCLUDE if not in array

Initialization:

4. Preference-Inference Document (COMPLETE)

File: working/ops/morning-briefing-redesign.md (Part 4)

12 Key Insights About Quan:

  1. Embodied vitality is non-negotiable (skiing, movement, recovery as state amplifier)

    • Evidence: Timeline refs, MEMORY.md, explicit weather request

  2. Charlie time protected absolutely (weekly block, pre-reset required)

    • Evidence: master-plan.md, relational entropy protocol, Friday scheduling

  3. Deep-work blocks are sacred (uninterruptible focus time)

    • Evidence: MEMORY.md founder energy constraint, 8 AM briefing move to protect early focus

  4. Interruptions created vs removed is weekly metric (net load matters more than output)

    • Evidence: MEMORY.md, system regression tracked, Quan's feedback on features adding load

  5. Ambiguity requires pushback, not assumption (ask, don't guess)

    • Evidence: Feb 9 explicit feedback "push back and ask for clarity"

  6. News must be actionable/contextual, not just reported (ZTAG relevance required)

    • Evidence: Engagement on specific intel vs generic headlines

  7. Tradeoff clarity > moral judgment (show the math, not "you should")

    • Evidence: MEMORY.md "optimize, don't opine" rule

  8. Charlie automation is force multiplier (Zoho Books + Xero pre-reconciliation)

    • Evidence: Feb 10 explicit decision "very high value way to help"

  9. Details matter (ZUES, Fathom links, weather) (consistency = trust building)

    • Evidence: Feb 10 corrections on spelling, meeting links, weather

  10. Cost transparency expected (daily tracking, not hidden in console)

    • Evidence: Feb 9 question about spend API, awareness of $13.50/day vs $5-7 target

  11. Async delivery > interruption (except truly urgent)

    • Evidence: Zapier → Drive workflow preference, Telegram announcements for briefing

  12. Master plan phases are foundational context (thinks in phases + milestones)

    • Evidence: Phase gates, quarterly milestones language, blocking factors tracked

Confidence Levels:

What Changed From Current Briefing

Current State Problems (Identified)

  1. ❌ Repeat news with no dedup (same articles cycle)
  2. ❌ No forest view (master plan invisible)
  3. ❌ No coherence check (Charlie time, entropy untracked)
  4. ❌ No sovereignty metric (interruptions not surfaced)
  5. ❌ Generic news noise (disconnected from ZTAG pipeline)
  6. ❌ Missing cost tracking (spend above target, invisible)
  7. ❌ Weather missing (explicitly requested, forgotten after one day)
  8. ❌ No deep-work protection (calendar integrated but not filtered)
  9. ❌ No decision support (ambiguities presented without pushback)
  10. ❌ Charlie automation invisible (high-priority initiative status hidden)

Redesigned Briefing Solutions

  1. ✅ news-shared.json tracking + dedup logic in cron prompt
  2. ✅ FOREST section shows phase + milestones + blockers + gate health
  3. ✅ COHERENCE CHECK section with reset scheduling, entropy assessment
  4. ✅ FOUNDER SOVEREIGNTY DASHBOARD with interruptions metric
  5. ✅ NEWS filtered to ZTAG relevance, impact category, action implications
  6. ✅ Cost tracking visible in OPERATIONAL CONTEXT (daily + target + trend)
  7. ✅ Weather always present (Castaic + Mammoth via wttr.in)
  8. ✅ Deep-work block status flagged (🟢 PROTECTED / 🟡 AT RISK / 🔴 LOST)
  9. ✅ DECISION POINTS section with ambiguity pushback + options
  10. ✅ Charlie automation in STRATEGIC INITIATIVES (status + blocking factors)

Files Created / Modified

Created (New)

To Be Modified (When Deploying)

To Be Created (By Quan/Minnie-Ops When Approved)

Deployment Checklist

Before 2026-02-11 08:00 AM PT (when new briefing should run):

After First Briefing (Feb 11, 8:00 AM):

Cost Impact

Current Morning Briefing:

Redesigned Briefing:

Net Daily Cost Change: +$0.10-0.20/day
Total Daily Impact (all agents): $13.50 → $13.60-$13.70 (within acceptable range for this phase)

Success Metrics (Post-Launch)

Technical (Must-Have):

Behavioral (Target):

Phase Gate (Critical):

Future Enhancement Opportunities

Out of scope for this redesign, but identified for Phase 5+:

  1. Personalized opportunity scoring

    • Surface ZTAG pipeline opportunities that fit Quan's travel calendar
    • Example: "Saugus RFP opens Mar 10. You'll be post-SHAPE (Mar 17+), full availability for follow-up."

  2. Relational entropy prediction

    • Model: "If Charlie-focused reset doesn't happen by [date], friction probability = Y%"
    • Proactive reset scheduling (not reactive to conflict)

  3. Deep-work protection

    • Automatically block calendar from scheduling over protected focus windows
    • Conflict resolution with reasoning (not just "conflict found")

  4. Cost anomaly detection

    • Alert if daily spend spikes 2x normal without explanation
    • Root cause analysis ("higher than usual because: news search found 20 results vs typical 5")

  5. Competitive move early warning

    • Flag mentions of Google/PowerSchool/IXL moves before mainstream news
    • Sourced from SEC filings, investor reports, job postings

  6. Phase gate automation

    • Suggest phase advance 3 days before gate closure
    • Milestone tracking with % complete indicators

  7. Charlie accounting dashboard

    • Real-time pre-reconciliation of Zoho Books + Xero with 1-click approvals
    • Expected impact: 10+ hours/month saved for Charlie (justifies Minnie investment alone)

Risks & Mitigation

Risk Likelihood Impact Mitigation
Weather API (wttr.in) downtime Low Briefing still delivers, weather section shows error gracefully Error handling in weather-fetch.py catches exceptions
News search spam (too many results) Medium Briefing becomes noise-heavy, loses signal Filter queries refined weekly based on result quality
Dedup logic misses repeats Low Quan sees duplicate news (regression) Manual review + adjustment of semantic similarity threshold
Cost increase above target Medium Phase budget constraint violated Monitor daily, optimize search queries + heartbeat interval
Reference/ files unpopulated Medium ZTAG context missing, news/pipeline less relevant Template works without reference/ (gracefully degrades), populate when available
Quantum changes preferences mid-way Low Briefing becomes misaligned Quick feedback loop (daily briefings), rapid iteration, MEMORY.md updates capture new preferences

What's Ready Now

Fully Operational

  1. ✅ Briefing template structure (markdown, sections, formatting)
  2. ✅ Prompt file (behavioral rules, dedup logic, edge cases)
  3. ✅ News dedup schema (JSON structure, categories, maintenance rules)
  4. ✅ Weather-fetch tool (wttr.in integration, error handling)
  5. ✅ Preference inference documentation (12 insights, confidence levels, behavioral rules)

Requires Quan Approval

  1. Redesign document review (structure, section ordering, content)
  2. Confirmation of reference/ file priorities (which to populate first)
  3. Approval of cron job changes (data-fetch steps, cost impact)

Requires Minnie-Ops Execution

  1. Extract and register cron/morning-briefing-prompt.md in cron jobs
  2. Update ~/.openclaw/cron/jobs.json with data-fetch steps
  3. First run and feedback collection (Feb 11 → Feb 15)
  4. Iteration cycles based on Quan's response

Next Steps for Main Agent

  1. Present redesign document to Quan
  2. Gather feedback on:
    • Template structure (missing sections? extra sections?)
    • News filtering criteria (what categories matter most?)
    • Cost impact (acceptable +$0.10-0.20/day?)
    • Implementation timing (deploy immediately? test first?)
  3. Approve changes to cron job configuration
  4. Execute deployment when Quan approves
  5. Monitor first 5 briefings for quality, iterate based on feedback
  6. Track success metrics (engagement rate, content accuracy, actionability)

Summary

Redesign scope: COMPLETE ✅

Delivered a strategic, loss-function-aligned morning briefing template that:

The redesign prevents repeat news, surfaces master plan context, flags coherence risks, and asks for clarity on ambiguities — all aligned to Quan's actual preferences inferred from 6,000+ lines of timeline and 10+ days of memory/feedback.

Status: Ready for deployment after Quan approval.

Created by: Minnie subagent (gradient descent optimizer)
Time invested: Strategic analysis + comprehensive documentation
Next: Awaiting main agent handoff to Quan for review + approval