Divorce Helper Subagent System

A collection of specialized Claude Code subagents designed to help individuals navigate the divorce process across legal, financial, parenting, and safety domains.

What This Does

This skill provides a multi-agent system where each subagent handles a specific aspect of divorce proceedings:

Instructions for the AI Agent

Project Structure

All subagents are stored in `.claude/agents/` as Markdown files with YAML frontmatter:

```markdown

---

name: agent-name

description: When and why to use this agent

tools: Read, Write, Bash, Grep, Glob

model: sonnet

---

System prompt for the agent...

```

Core Data Structures

Use these three data structures across all subagents:

1. **Vault** - Encrypted store for IDs, finances, kids' records

2. **Timeline** - Dated log of events, orders, communications

3. **Red Flags** - Violence, relocation, asset flight, coercive control, substance abuse indicators

Legal & Ethical Guardrails (CRITICAL)

**NEVER advise or implement:**

**ALWAYS enforce:**

Divorce Process Flow

Guide users through this sequence:

1. **Safety & Stabilization** (Days 0-7) - Safety planning, protective orders, digital hygiene

2. **Resolution Track Selection** - Mediation, collaborative divorce, or litigation

3. **Pre-Filing Prep** (Weeks 1-3) - Document collection, credit checks, budget creation

4. **Filing & Service** - Petition filing, service, temporary restraints

5. **Temporary Orders** (30-90 days) - Parenting schedule, support, status quo

6. **Financial Disclosures & Discovery** - Mandatory disclosures, evidence preservation

7. **Parenting Plan Development** - Best interests analysis, UCCJEA compliance

8. **Property Division** - Characterization, valuation, division

9. **Support Determination** - Child support (guideline), spousal support (factor-based)

10. **Settlement or Trial** - Proposals, MSA/parenting plan, or trial prep

11. **Judgment & Post-Decree** - Retitling, QDRO filing, beneficiary updates

Jurisdictional Awareness

**Adapt guidance based on:**

Creating New Subagents

When creating subagents:

1. Use `/agents` command for guided creation

2. Store in `.claude/agents/` as Markdown with YAML frontmatter

3. Define clear, non-overlapping responsibilities

4. Specify required tools (Read, Write, Bash, Grep, Glob)

5. Build legal/ethical guardrails directly into system prompt

6. Follow BINF communication principle (Brief, Informative, Neutral, Firm)

Subagent Invocation

View available agents with `/agents` command. Claude Code will automatically delegate based on task description, or you can explicitly request an agent by name.

Security Requirements

Implement these security measures:

Key External Resources

Reference these resources for jurisdiction-specific guidance:

Development Priorities

When modifying or creating subagents:

1. **Safety First** - Prioritize user safety and legal compliance

2. **Evidence Preservation** - Never encourage deletion/hiding

3. **Jurisdiction Awareness** - Adapt to user location

4. **Child-Centric** - Focus on best interests of children

5. **Clear Boundaries** - Focused, non-overlapping responsibilities

6. **Proactive Guardrails** - Built-in legal/ethical constraints

Example Usage

```bash

View available subagents

/agents

User asks: "My spouse emptied our joint account yesterday"

→ Delegate to Financial Forensics & Safety agents

User asks: "We're in different states, where do we file custody?"

→ Delegate to Co-Parenting agent (UCCJEA compliance)

User asks: "How do I gather evidence of hidden income?"

→ Delegate to Document & Evidence + Financial Forensics agents

```

Important Notes