Claude-skill-registry brainstorming-exploring

The EXPLORING phase gathers context about the project and forms an initial understanding

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/brainstorming-exploring" ~/.claude/skills/majiayu000-claude-skill-registry-brainstorming-exploring && rm -rf "$T"
manifest: skills/data/brainstorming-exploring/SKILL.md
source content

EXPLORING Phase

The EXPLORING phase gathers context about the project and forms an initial understanding of what needs to be built.

Purpose

  • Understand the current project state
  • Read relevant files and documentation
  • Check recent git history for context
  • Form an initial list of items/topics to discuss

Step 0: Query Kodex

Before reading files, check project knowledge base for relevant context.

Topic Inference

  1. Get current work item from collab-state.json
  2. Extract keywords from item title/description
  3. Build topic candidates:
    • {keyword}
    • {keyword}-patterns
    • {keyword}-conventions

Query Process

FOR each candidate topic name:
  Tool: mcp__plugin_mermaid-collab_mermaid__kodex_query_topic
  Args: { "project": "<cwd>", "name": "<candidate>" }

  IF found: Add to context

Example

For work item "Add user authentication":

Tool: mcp__plugin_mermaid-collab_mermaid__kodex_query_topic
Args: {
  "project": "<absolute-path-to-cwd>",
  "name": "authentication"
}

Fallback

If no topics found from title keywords:

  1. Try keywords from item description
  2. Try removing suffixes (-patterns, -conventions)
  3. Continue to file exploration without Kodex context

Display found topics before proceeding to "Check project state".

Process

  1. Check project state:

    • Read relevant files (code, docs, configs)
    • Check recent commits for context
    • Understand existing architecture
  2. Gather requirements context:

    • Review any existing specs or requirements
    • Understand stakeholder expectations
    • Identify constraints and dependencies
  3. Form initial item list:

    • Create a preliminary list of topics to discuss
    • Group related items together
    • Prioritize by importance/dependencies

Single-Item Mode (EXPLORING)

When

currentItem
is set in collab-state.json:

  • Read relevant files based on item description
  • Check git history for related changes
  • Gather context specific to this item only
  • Do NOT expand scope beyond this item

Exit Criteria

  • Context gathered from relevant files
  • Initial list of items/topics formed
  • Ready to discuss items one at a time

Checkpoint: Current State Diagram

REQUIRED before proceeding to CLARIFYING:

Create a diagram showing the current state relevant to this work:

Tool: mcp__plugin_mermaid-collab_mermaid__create_diagram
Args: {
  "project": "<cwd>",
  "session": "<session>",
  "name": "current-state-item-N",
  "content": <flowchart showing existing components/flow>
}

Display: "Current state diagram: [previewUrl]"

If the work item doesn't involve architecture/flow (e.g., pure text changes), create a simple diagram showing the file(s) being modified:

graph LR
    A[File: path/to/file.md] --> B[Section being modified]

Transition to CLARIFYING

Prerequisites:

  • Read relevant files/context
  • Formed initial list of items

Announce: "I've gathered context. Now let me discuss each item with you one at a time."

Completion

At the end of this skill's work, call complete_skill:

Tool: mcp__plugin_mermaid-collab_mermaid__complete_skill
Args: { "project": "<cwd>", "session": "<session>", "skill": "brainstorming-exploring" }

Handle response:

  • If
    action == "clear"
    : Invoke skill: collab-clear
  • If
    next_skill
    is not null: Invoke that skill
  • If
    next_skill
    is null: Workflow complete