danieldekay/ai-ops/project-planning

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Strategic project planning system for creating and maintaining planning artifacts. Use when creating or updating vision, roadmap, epics, backlog, project status dashboards, or milestone reviews. Use when breaking down large goals into actionable work. Use when triaging scope (L1 spec / L2 task-batch / L3 quick-fix). Templates live in this skill's templates/ directory. NOT for spec-kit artifacts (spec.md, plan.md, tasks.md) — those belong to speckit agents.

Install

npx vskill@latest install danieldekay/ai-ops/project-planning

SKILL.md

Project Planning Skill

Strategic planning system that produces living planning documents. Covers everything ABOVE the spec level: vision, roadmap, epics, backlog, status dashboards, and milestone reviews.

When To Use This Skill

  • Creating a new project's planning structure from scratch
  • Writing or updating a vision, roadmap, epic, backlog, or status dashboard
  • Running a milestone review or checkpoint
  • Triaging work into L1 (spec) / L2 (task-batch) / L3 (quick-fix)
  • Extracting deferred items from specs into a centralized backlog

When NOT To Use This Skill

  • Writing spec.md, plan.md, tasks.md → use speckit agents
  • Implementing code → use spec-implementer or direct coding
  • Reviewing code → use Code Review Agent

The Planning Hierarchy

VISION → EPICS → WORK ITEMS → EXECUTION
↓ ↓ ↓
L1 L2 L3
Spec Batch Fix

Every piece of work enters the hierarchy and gets triaged by size before execution.

Artifact Catalog

ArtifactTemplateWhen To CreateUpdate Frequency
Planning Sessiontemplates/planning-session.mdEvery planner sessionOnce (immutable snapshot)
Visiontemplates/vision.mdOnce per projectRarely (strategy shifts)
Roadmaptemplates/roadmap.mdOnce per projectWhen priorities change
Epictemplates/epic.mdPer capability group (5–15 specs)When epic scope changes
Project Statustemplates/project-status.mdOnce per projectAfter any spec work
Backlogtemplates/backlog.mdOnce per projectWhen completing or discovering items
Milestone Reviewtemplates/milestone-review.mdAt each milestone boundaryOnce (snapshot)
Using Templates
  1. Read the template from templates/<name>.md
  2. Replace all {{PLACEHOLDER}} values with project-specific content
  3. Write to the target location (see File Locations below)
  4. Remove any template sections that don't apply
Planning Session Artifacts (Mandatory)

Every planner session MUST produce a dated session artifact at:

plan/sessions/YYYY-MM-DD-<topic-slug>.md

Examples:

  • plan/sessions/2026-03-02-situation-assessment.md
  • plan/sessions/2026-03-02-epic-05-breakdown.md
  • plan/sessions/2026-03-02-backlog-extraction.md

This creates a searchable history of planning decisions even when the main artifacts (ROADMAP, BACKLOG) are updated in-place.

Work Item Triage (L1 / L2 / L3)

See references/triage-rules.md for the complete decision tree.

Quick version:

  • L1 Spec: New domain entities, cross-layer changes, 3+ user stories → speckit.specify
  • L2 Task-Batch: Clear scope, no new domain concepts, 1–3 sessions → entry in backlog
  • L3 Quick-Fix: Single file, <30 min, near-zero risk → just do it

When uncertain, default to L2. You can promote to L1 if complexity emerges. L1 → L2 is rare and indicates overscoping.

Backlog Management

Extracting Deferred Items

To build or refresh a backlog:

  1. Scan all specs/*/future-improvements.md files
  2. Scan plan/MVP-CHECKLIST.md for open blockers
  3. Check each spec for incomplete phases or user stories
  4. Consolidate into plan/BACKLOG.md using the backlog template
  5. Assign IDs: BLK-NNN (blocker), RDY-NNN (ready), SCP-NNN (needs scoping), DEF-NNN (deferred)
  6. Assign effort level: L1, L2, or L3
Prioritization Framework
  1. Blockers — anything preventing other work
  2. Current branch — finish before starting new
  3. Critical path — items on the path to launch
  4. Multipliers — features that unlock many others
  5. Quick wins — high value, low effort
  6. Nice-to-haves — someday/maybe

Creating New Epics

Epics group 5–15 related specs into a coherent capability. Use templates/epic.md.

Rules:

  • Every spec MUST belong to exactly one epic
  • Epic names describe a capability, not a feature (e.g., "Entity Expansion" not "Add People")
  • Include dependency links to other epics
  • Track completion as aggregate of contained spec statuses

Milestone Reviews

At milestone boundaries, create a review using templates/milestone-review.md:

  1. What was planned vs. delivered
  2. What slipped and why
  3. Backlog items discovered during the milestone
  4. Scope changes for the next milestone
  5. Updated risk assessment

File Locations

Planning artifacts live in plan/:

plan/
├── WORKFLOW.md # How the planning system works (reference)
├── PROJECT-STATUS.md # Spec-level dashboard
├── ROADMAP.md # Milestones and execution lanes
├── BACKLOG.md # Deferred work items
├── sessions/ # Dated per-session artifacts (one per planner run)
│ └── YYYY-MM-DD-*.md
├── northstar/
│ ├── VISION.md # Strategic product truth
│ ├── TAXONOMY.md # Domain model definitions
│ ├── epics/ # Epic definitions
│ │ ├── README.md # Epic index
│ │ └── EPIC-NN-*.md # Individual epics
│ ├── personas/ # User archetypes
│ └── references/ # Brand, migration docs

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/danieldekay/ai-ops/project-planning/badge)](https://verified-skill.com/skills/danieldekay/ai-ops/project-planning)

Something wrong with this skill or its page? Tell us.