luokai0/ai-agent-skills-by-luo-kai/meeting-to-plan-integrator

Trust tier 3
  • community
Source check pending

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

Takes decisions from a meeting (sprint demo, retro, stakeholder call, team sync) and applies them to the release plan and related artifacts. The 'decisions become actions' bridge. Use when asked to: apply meeting decisions, update the plan from meeting notes, integrate feedback, or cascade changes from a meeting.

Install

npx vskill@latest install luokai0/ai-agent-skills-by-luo-kai/meeting-to-plan-integrator

SKILL.md

Meeting-to-Plan Integrator — Entry Orchestrator

NON-NEGOTIABLE (read first)

  1. Save every artifact to file before presenting in chat. Chat is ephemeral.
  2. Never fabricate decisions. Every change must trace to a specific decision from the meeting. Unknown = [TBD].
  3. Never skip checkpoints. Every CHECKPOINT requires user confirmation before proceeding.
  4. Graded provenance: every edit to a target document carries source attribution: (Source: [meeting-name], [date]). Decisions are authoritative claims. Impact analysis is synthesized analysis. Cascade predictions are inference and must be confirmed before applying.
  5. One stage at a time. Complete and save before starting the next.
  6. Never apply changes the user hasn't confirmed. The change manifest is the contract.

Purpose

Take decisions from a meeting and apply them to the release plan and related artifacts. This is the "decisions become actions" bridge.

What you'll produce:

  • Updated target documents with source-attributed edits
  • Change-Manifest-[date].md — record of what changed, why, and where

When to invoke:

  • After a sprint demo where stakeholders gave feedback
  • After a retrospective with action items
  • After a stakeholder call with scope or timeline changes
  • After any meeting that produced decisions affecting the plan

Workflow Chain

[You are here]
SKILL.md (Intake)
↓ reads
workflows/01-intake.md → loads meeting decisions, identifies targets
↓ reads
workflows/02-extract-changes.md → classifies decisions, builds change manifest
↓ reads
workflows/03-apply-changes.md → applies confirmed changes to documents
↓ reads
workflows/04-cascade.md → checks for downstream effects

Each workflow file is read in turn. Do NOT skip ahead.


Step 1: Read Shared Context

MANDATORY READ the following files before proceeding:

Read the file:

../shared/delivery-model.md

Read the file:

../shared/execution-rules.md


Step 2: Start Intake

Read the file:

workflows/01-intake.md

Follow that file's instructions completely. When complete, return here and proceed to Step 3.


Step 3: Extract Changes

Read the file:

workflows/02-extract-changes.md

Follow that file's instructions completely. When complete, return here and proceed to Step 4.


Step 4: Apply Changes

Read the file:

workflows/03-apply-changes.md

Follow that file's instructions completely. When complete, return here and proceed to Step 5.


Step 5: Cascade Check

Read the file:

workflows/04-cascade.md

Follow that file's instructions completely. Skill is done.


Related Skills

SituationSkill
Generate meeting notes from a transcript (input to this skill)transcript-to-meeting-notes
Create or update the release planrelease-sprint-planner
Plan a sprint kickoffsprint-planning-session
Track sprint progresssprint-progress-tracker
Generate sprint reviewsprint-review-generator

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/luokai0/ai-agent-skills-by-luo-kai/meeting-to-plan-integrator/badge)](https://verified-skill.com/skills/luokai0/ai-agent-skills-by-luo-kai/meeting-to-plan-integrator)

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