Skip to content

Daily Plan

AI-assisted daily planning workspace for Wenbing.

The goal is to prepare a focused next-day morning brief so Wenbing can wake up and execute instead of spending the morning deciding what matters.

Canonical loop

  1. Nightly generation
  2. Agent reads agent/goal.md, state/, recent briefs/*/human-feedback.md, recent evening-review.md, and live/current sources.
  3. Agent writes briefs/YYYY-MM-DD/ for the next day.
  4. Morning execution
  5. Wenbing opens briefs/YYYY-MM-DD/MORNING-BRIEF.md.
  6. MORNING-BRIEF.md is the only canonical accepted plan.
  7. Feedback / correction
  8. Wenbing writes feedback in human-feedback.md or chat.
  9. If a revised plan is accepted, promote it back to MORNING-BRIEF.md, archive older versions, and regenerate evening-review.md.
  10. End-of-day review
  11. Wenbing fills evening-review.md quickly.
  12. Agent updates state/open-items.md, state/decisions.md, and eventually state/learnings.md.

Folder layout

daily-plan/
├── agent/
│   └── goal.md                  # Nightly agent instructions
├── state/
│   ├── preferences.md           # Stable planning preferences and corrected assumptions
│   ├── open-items.md            # Rolling carry-over queue
│   ├── recurring.md             # Required daily/weekly/monthly checks
│   ├── decisions.md             # Explicitly dropped/deferred items
│   └── learnings.md             # Repeated patterns distilled over time
├── briefs/
│   └── YYYY-MM-DD/
│       ├── MORNING-BRIEF.md     # Canonical accepted plan for that day
│       ├── MORNING-BRIEF-vN.md  # Optional revisions / archived versions
│       ├── README.md            # Day packet index
│       ├── reports/             # Supporting reports/action kits
│       ├── analysis/            # Evidence audit/source ledger
│       ├── raw/                 # Raw snapshots
│       ├── human-feedback.md    # Wenbing's feedback
│       └── evening-review.md    # Planned vs actual
└── README.md

Source priority for generating plans

  1. Active sprint Jira commitments.
  2. Direct manager/scrum expectations and hard deadlines.
  3. Direct Teams/email asks mentioning Wenbing.
  4. Personal/life urgent lane.
  5. Oncall/incidents only when Wenbing owns/is oncall/is directly escalated.
  6. Strategic but non-urgent work.
  7. Parking-lot cleanup.

Required active sprint query

Every brief near sprint end must run or clearly emulate:

assignee = currentUser()
AND sprint in openSprints()
AND resolution is EMPTY
ORDER BY priority ASC, status ASC

Then show separate sections for:

  • Active sprint commitments
  • Assigned but not active sprint
  • Reported-by-me follow-ups
  • Parking lot / explicit non-priorities

Acceptance rules

  • If the generated plan is revised after feedback, do not leave the accepted plan only as MORNING-BRIEF-v2.md.
  • Promote accepted content to MORNING-BRIEF.md.
  • Preserve rejected/older plans as MORNING-BRIEF-v1.md, MORNING-BRIEF-v2.md, etc.
  • Update the day README.md and evening-review.md to match the accepted plan.

Current scheduled automation

A nightly cron job should generate the next-day plan and deliver a concise summary back to Telegram. Check Hermes cron jobs for the exact schedule and last run status.