Daily Workflow Skill (English Version)

Manage daily work sessions with automated project documentation tracking and AI-to-AI handover support.

When to Use This Skill

Trigger this skill when the user mentions:

• "starting work" / "start of day" / "let's begin" - Starting the work day
• "ending work" / "end of day" / "wrapping up" - Ending the work day

These phrases indicate intentional work session boundaries where project state should be captured and restored.

Core Workflow

Phase 1: Start Work Session

Execute the following sequence when the user indicates work is starting:

1. Ensure Docs Directory Exists

   • Check if Docs/ directory exists in workspace root
   • Create Docs/ if it does not exist

2. Check and Create Required Files

   • Verify these five files exist in Docs/:
     • PROJECT_TARGET.md - Project objectives and goals
     • PROJECT_STATUS.md - Current project status
     • COMPLETED_JOBS.md - Completed work items
     • PENDING_JOBS.md - Pending/incomplete work items
     • NEXT_STEPS.md - Planned next steps
   • Create any missing files with appropriate template headers

3. Read All Documentation

   • Read all five files from Docs/
   • Extract key information:
     • Current project objectives
     • Latest project status
     • Recently completed work
     • Pending tasks and blockers
     • Planned next steps

4. Present Work Session Briefing

   • Summarize current project state
   • Highlight completed work since last session
   • Identify pending tasks
   • Suggest priority work for current session
   • Ask user for confirmation or adjustments to the plan

Phase 2: End Work Session

Execute the following sequence when the user indicates work is ending:

1. Update Completed Jobs

   • Append today's completed work items to Docs/COMPLETED_JOBS.md
   • Use format: ## [YYYY-MM-DD]\n- [completed item 1]\n- [completed item 2]
   • Include sufficient detail for another AI to understand what was done

2. Update Pending Jobs

   • Update Docs/PENDING_JOBS.md with:
     • Tasks started but not completed
     • Blockers encountered
     • Tasks deferred to next session
   • Remove completed items from pending list

3. Update Project Status

   • Update Docs/PROJECT_STATUS.md with:
     • Current completion percentage
     • Key milestones reached
     • Current focus area
     • Any changes in project direction

4. Update Project Target (if needed)

   • Modify Docs/PROJECT_TARGET.md only if:
     • Project objectives have changed
     • New requirements discovered
     • Scope adjustments needed
   • Otherwise, leave unchanged

5. Update Next Steps

   • Write clear, actionable next steps to Docs/NEXT_STEPS.md
   • Include:
     • Specific tasks to tackle next session
     • Priority order
     • Any prerequisites or dependencies
     • Context needed to resume work immediately

6. Present Session Summary

   • Summarize what was completed
   • List updated documentation files
   • Confirm next steps are clearly documented
   • Ensure handover-ready state for next AI

File Templates

When creating missing files, use these templates:

PROJECT_TARGET.md

# Project Target

## Project Overview
[Describe the project's main objective]

## Key Goals
- [Goal 1]
- [Goal 2]

## Success Criteria
- [Criterion 1]
- [Criterion 2]

## Last Updated
[YYYY-MM-DD]

PROJECT_STATUS.md

# Project Status

## Current Status
[Brief description of current state]

## Completion
[XX]% complete

## Current Focus
[What is being worked on now]

## Recent Milestones
- [YYYY-MM-DD] [Milestone description]

## Last Updated
[YYYY-MM-DD]

COMPLETED_JOBS.md

# Completed Jobs

## [YYYY-MM-DD]
- [Completed task 1]
- [Completed task 2]

## [YYYY-MM-DD]
- [Completed task 1]

PENDING_JOBS.md

# Pending Jobs

## High Priority
- [ ] [Task 1]
- [ ] [Task 2]

## Medium Priority
- [ ] [Task 3]

## Low Priority
- [ ] [Task 4]

## Blockers
- [Blocker description if any]

NEXT_STEPS.md

# Next Steps

## Immediate Actions (Next Session)
1. [Action 1 - with context]
2. [Action 2 - with context]

## Upcoming Tasks
- [Task 1]
- [Task 2]

## Notes for Next AI
[Important context, decisions made, things to remember]

AI-to-AI Handover Principles

Write all documentation assuming the next reader will be a different AI instance that needs to:

• Understand what was being built and why
• Resume work without asking basic questions
• Continue the same coding style and conventions
• Respect decisions already made

Key practices:

• Write in detail, not shorthand
• Explain the "why" not just the "what"
• Include code snippets or file references when relevant
• Note any workarounds, hacks, or technical debt
• Record user preferences and decisions

Workflow Diagram

User indicates work starting
    ↓
Check Docs/ exists → Create if missing
    ↓
Check 5 files exist → Create missing with templates
    ↓
Read all 5 files
    ↓
Present work session briefing
    ↓
User works with AI assistance
    ↓
User indicates work ending
    ↓
Update COMPLETED_JOBS.md
    ↓
Update PENDING_JOBS.md
    ↓
Update PROJECT_STATUS.md
    ↓
Update PROJECT_TARGET.md (if needed)
    ↓
Update NEXT_STEPS.md
    ↓
Present session summary
    ↓
Project state saved for next AI

Important Notes

• Always use absolute paths when referencing Docs/ directory
• Append to COMPLETED_JOBS.md (never overwrite previous entries)
• Overwrite NEXT_STEPS.md each session (it's for the immediate next session)
• Preserve historical information in PROJECT_STATUS.md and COMPLETED_JOBS.md
• When in doubt, write more context, not less
• This skill supports both English and Chinese trigger phrases for flexibility