AgentSkillsCN

Museum

配合博物馆开发追踪系统

SKILL.md
--- frontmatter
description: Work with museum development tracking system
argument-hint: "[session|capture|create|link|search|<id>]"

Museum Development Tracking System

Track creative worldbuilding decisions, brainstorming sessions, and design evolution with full traceability.

Usage

  • /museum - List all museum development items
  • /museum session "Title" - Start a new brainstorming session
  • /museum capture <sessionId> decision "Content" - Capture a decision during session
  • /museum capture <sessionId> question "Content" - Capture a question during session
  • /museum <id> - View item details
  • /museum <id> verdict accepted "Rationale" - Set verdict on a decision
  • /museum link <from> <to> spawned "Note" - Link two items
  • /museum search "query" - Search items

Arguments

$ARGUMENTS - Command and arguments (see help for full list)

Instructions

For no arguments or "list":

Show the current museum development tracker status:

bash
node scripts/museum-dev.js list

For "session <title>":

Start a new brainstorming session:

bash
node scripts/museum-dev.js session "Session Title"

This creates a session item and returns its ID. Use this ID with capture commands to link decisions/questions to the session.

For "capture <sessionId> <type> <content>":

Quickly capture an item during a brainstorming session:

bash
node scripts/museum-dev.js capture <sessionId> decision "The Order are tragic protectors, not villains"
node scripts/museum-dev.js capture <sessionId> question "Why hasn't the Order silenced Austen?"

Valid types: session, decision, question, element, reference

The captured item is automatically linked to the session with a "spawned" relationship.

For "session-end <sessionId>":

End a session and optionally attach the transcript:

bash
node scripts/museum-dev.js session-end <sessionId>
node scripts/museum-dev.js session-end <sessionId> --transcript ./transcript.md

For "create <type> <title>":

Create a standalone item (not linked to a session):

bash
node scripts/museum-dev.js create decision "Title"
node scripts/museum-dev.js create question "Title"
node scripts/museum-dev.js create element "Title" --element-type wing
node scripts/museum-dev.js create reference "Title"

Element subtypes: wing, artifact, plaque, character, mechanic, audio, visual

For "<id>" (viewing an item):

View full details of an item:

bash
node scripts/museum-dev.js <id>

Partial IDs (8+ characters) are resolved automatically.

For "<id> <status>":

Update item status:

bash
node scripts/museum-dev.js <id> in-progress "Starting work"
node scripts/museum-dev.js <id> completed "Documented in museum-layout.md"

Valid statuses: new, in-progress, in-review, completed, archived

For "<id> verdict <type> <rationale>":

Set a verdict on a decision:

bash
node scripts/museum-dev.js <id> verdict accepted "Aligns with the protective tragedy theme"
node scripts/museum-dev.js <id> verdict rejected "Too on-the-nose, breaks immersion"
node scripts/museum-dev.js <id> verdict deferred "Revisit after Phase 2"
node scripts/museum-dev.js <id> verdict superseded "Replaced by decision XYZ"
  • accepted moves the item to completed
  • rejected archives the item but preserves it for historical context
  • deferred keeps it for later consideration
  • superseded indicates it was replaced by a better idea

For "<id> answer <answer>":

Answer a question item:

bash
node scripts/museum-dev.js <id> answer "The Order hasn't silenced Austen because..."

For "link <from> <to> <type> [note]":

Create a bidirectional link between items:

bash
node scripts/museum-dev.js link <sessionId> <decisionId> spawned "From brainstorm"
node scripts/museum-dev.js link <id1> <id2> derived "Evolved from"
node scripts/museum-dev.js link <id1> <id2> related "See also"
node scripts/museum-dev.js link <id1> <id2> contradicts "Conflicts with"
node scripts/museum-dev.js link <questionId> <decisionId> answers "This resolves the question"

Link types: spawned, derived, related, contradicts, answers

For "links <id>":

Show all links to/from an item:

bash
node scripts/museum-dev.js links <id>

For "trace <id>":

Trace an item back to its source session(s):

bash
node scripts/museum-dev.js trace <id>

For "tree <sessionId>":

Show all items spawned from a session:

bash
node scripts/museum-dev.js tree <sessionId>

For "search <query>":

Search across all items:

bash
node scripts/museum-dev.js search "order"
node scripts/museum-dev.js search "vtg"

For "journal <id>":

View the activity journal for an item:

bash
node scripts/museum-dev.js journal <id>

For "transcript <sessionId>":

View the full transcript attached to a session:

bash
node scripts/museum-dev.js transcript <sessionId>

For tags and attachments:

bash
node scripts/museum-dev.js <id> tag add order
node scripts/museum-dev.js <id> tag remove vtg
node scripts/museum-dev.js <id> attach ./concept-art.png "Early Order concept"
node scripts/museum-dev.js <id> attach-url https://... "Reference material"
node scripts/museum-dev.js <id> attachments

For help:

bash
node scripts/museum-dev.js help

Workflow Example

  1. Start a session:

    bash
    node scripts/museum-dev.js session "Order motivation brainstorm"
    # Returns: SESSION_ID=abc123xyz
    
  2. Capture ideas during conversation:

    bash
    node scripts/museum-dev.js capture abc123 decision "Order = tragic protectors, not villains"
    node scripts/museum-dev.js capture abc123 question "Why hasn't Order silenced Austen?"
    node scripts/museum-dev.js capture abc123 decision "VTG origin story - Oakland fire spinners"
    
  3. End session with transcript:

    bash
    node scripts/museum-dev.js session-end abc123 --transcript ./session-transcript.md
    
  4. Set verdicts later:

    bash
    node scripts/museum-dev.js <decisionId> verdict accepted "Core to the narrative"
    
  5. Trace any decision back to its source:

    bash
    node scripts/museum-dev.js trace <decisionId>
    

Key Principles

  1. Capture everything - rejected ideas are as valuable as accepted ones for understanding the creative process
  2. Link liberally - connections help trace the evolution of ideas
  3. Store transcripts - raw conversations are more valuable than summaries
  4. Set verdicts - accepted/rejected/deferred helps filter but preserves history
  5. Trace back - any decision should be traceable to the session that spawned it