Onboard Milestones

SkillDev tools

Create MILESTONES.md from existing workflow docs — scan, propose, refine, write.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Onboard Milestones skill

What this skill tells your AI

The instructions your AI receives, as published by neuron-mr-white/unipi in packages/milestone/skills/milestone-onboard/SKILL.md and read by ahel’s review.

Create a MILESTONES.md file by scanning existing workflow documentation. Groups scattered tasks into coherent milestone phases.

Boundaries

This skill MAY: read .unipi/docs/, read existing specs/plans/quick-work/debug/fix/chore docs, write .unipi/docs/MILESTONES.md. This skill MAY NOT: modify existing workflow docs, delete files, merge branches.


Phase 1: Explore

Scan .unipi/docs/ for existing workflow documentation. Understand what's been done and what's planned.

  1. List all files in .unipi/docs/{specs,plans,quick-work,debug,fix,chore}/
  2. For each file, extract:
    • Checkbox items (- [ ] and - [x])
    • Task statuses (unstarted:, in-progress:, completed:)
    • File modification dates
  3. Categorize findings:
    • Completed work — checked items, completed tasks
    • In progress — in-progress tasks, partially checked lists
    • Planned — unstarted tasks, unchecked items
  4. Present summary to user:

    "Found X completed items, Y in-progress, Z planned across N documents."


Phase 2: Propose

Group findings into logical milestone phases. Present with rationale.

  1. Analyze themes across documents — group related items together

  2. Suggest 2-5 phases with clear names and descriptions

  3. For each phase, list proposed items (both done and todo)

  4. Present proposal:

    "Phase 1: Foundation (3/5 done)

    • Project scaffold
    • Core parser
    • Type definitions
    • Error handling
    • Documentation

    Phase 2: Features (0/3 done)

    • User dashboard
    • File upload
    • Notifications

    Does this grouping look right?"

  5. One question at a time — ask if phases are correct before proceeding


Phase 3: Refine

User approves/adjusts phases. Iterate until satisfied.

  1. If user wants changes:
    • Add phase: Ask for name and items
    • Remove phase: Confirm removal
    • Move items: Ask which item, which phase
    • Rename phase: Ask for new name
    • Add items: Ask for text and target phase
  2. After each change, show updated proposal
  3. Continue until user says "looks good" or "write it"

Phase 4: Write

Save MILESTONES.md using the milestone parser.

  1. Build MilestoneDoc from approved phases:
    • title: "Project Milestones" (or ask user for custom title)
    • created: today's date
    • updated: today's date
    • phases: approved phases with items
  2. Call writeMilestones(".unipi/docs/MILESTONES.md", doc)
  3. Verify file was written correctly

Phase 5: Report

Show summary and suggest next steps.

  1. Display what was written:

    "✅ MILESTONES.md created with N phases and M items.

    • Phase 1: Foundation (3/5 done)
    • Phase 2: Features (0/3 done)

    Next steps:

    • /unipi:milestone-update — sync milestones with completed work
    • Milestones will auto-inject context on session start
    • Completed items auto-sync on session end"

Validation Checklist

Before completing, verify:

  • MILESTONES.md exists at .unipi/docs/MILESTONES.md
  • File has valid frontmatter (title, created, updated)
  • All phases have names
  • All items have checkbox format (- [ ] or - [x])
  • Previously completed items are marked [x]
  • File parses correctly via parseMilestones()

Signals

GitHub stars
64
Forks
15
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
milestone-onboard
Source
github.com/neuron-mr-white/unipi