Implement LLM-driven governance architecture with structured memory

This commit completes the transition to a pure LLM-driven agentic
governance system with no hard-coded governance logic.

Core Architecture Changes:
- Add structured memory system (memory.py) for tracking governance processes
- Add LLM tools (tools.py) for deterministic operations (math, dates, random)
- Add audit trail system (audit.py) for human-readable decision explanations
- Add LLM-driven agent (agent_refactored.py) that interprets constitution

Documentation:
- Add ARCHITECTURE.md describing process-centric design
- Add ARCHITECTURE_EXAMPLE.md with complete workflow walkthrough
- Update README.md to reflect current LLM-driven architecture
- Simplify constitution.md to benevolent dictator model for testing

Templates:
- Add 8 governance templates (petition, consensus, do-ocracy, jury, etc.)
- Add 8 dispute resolution templates
- All templates work with generic process-based architecture

Key Design Principles:
- "Process" is central abstraction (not "proposal")
- No hard-coded process types or thresholds
- LLM interprets constitution to understand governance rules
- Tools ensure correctness for calculations
- Complete auditability with reasoning and citations

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
This commit is contained in:
Nathan Schneider
2026-02-08 14:24:23 -07:00
parent 5fe22060e1
commit bda868cb45
26 changed files with 8683 additions and 187 deletions

View File

@@ -0,0 +1,343 @@
# Dispute Resolution Templates
This directory contains comprehensive dispute resolution protocols that can be integrated into governance constitutions. Each template provides a complete framework for handling conflicts and disputes in online and offline communities.
## Overview
Effective governance requires not just decision-making processes, but also ways to handle conflicts when they arise. These templates are based on proven dispute resolution models from the [Community Rule Builder project](https://git.medlab.host/dispute-protocol/builder-prototype) and adapted for use with agentic governance bots.
## How to Use These Templates
### Option 1: Standalone Integration
Add a dispute resolution article to your existing constitution:
1. Choose a dispute resolution template
2. Copy it into your constitution as a new article (e.g., "Article 7: Dispute Resolution")
3. Adjust details to fit your community context
4. Configure your bot to recognize and facilitate the process
### Option 2: Multiple Pathways
Implement several processes for different situations:
- Peer-to-peer for minor conflicts
- Mediation or circles for deeper issues
- Jury or referee for formal decisions
- Different intensities for different needs
### Option 3: Escalation Ladder
Create a progression of dispute resolution:
1. Start with peer-to-peer
2. Escalate to chosen facilitator if needed
3. Move to circles or mediation for deeper work
4. Use jury/referee for binding decisions
5. Each level more structured and formal
## Available Templates
### 1. Peer-to-Peer (`peer-to-peer.md`)
**"A self-facilitated process where participants work together directly to resolve disputes"**
**Best for:**
- Minor interpersonal conflicts
- Misunderstandings needing clarification
- Communities valuing member autonomy
- First step in escalation ladder
**Key features:**
- No third-party facilitator
- Self-managed by parties
- Voluntary participation
- Simple ground rules
- Quick resolution
- Escalation path available
**Complexity:** Low | **Time:** Hours to days | **Formality:** Minimal
---
### 2. Chosen Facilitator (`chosen-facilitator.md`)
**"Participants mutually select a facilitator to help guide their dispute resolution"**
**Best for:**
- Conflicts needing structured support
- Situations where trust in process is crucial
- Cases requiring neutral guidance
- Communities with trained facilitators
**Key features:**
- Parties jointly select facilitator
- Structured process with ground rules
- Facilitator guides but doesn't decide
- 1-3 sessions typically
- Support persons allowed
- Clear appeal process
**Complexity:** Medium | **Time:** 1-3 weeks | **Formality:** Medium
---
### 3. Restorative Justice (`restorative-justice.md`)
**"A collaborative process emphasizing healing relationships and addressing harm through community engagement"**
**Best for:**
- Harm between community members
- Trust breakdowns requiring repair
- Situations needing community support
- Focus on healing over punishment
**Key features:**
- Circle process with talking piece
- Trained circle keeper
- Community participation
- Consensus-based agreements
- Accountability through repair
- Focus on transformation
**Complexity:** Medium-High | **Time:** Weeks to months | **Formality:** Medium
---
### 4. Transformative Justice (`transformative-justice.md`)
**"A process addressing immediate harm while transforming conditions that enabled it"**
**Best for:**
- Complex harm requiring deep work
- Addressing systemic issues
- Pattern behavior needing transformation
- Alternatives to state intervention
**Key features:**
- Multi-level analysis (individual to systemic)
- Support teams for all parties
- Long-term commitment
- Root cause focus
- Community education
- Liberation-oriented
**Complexity:** High | **Time:** Months to years | **Formality:** Structured but flexible
---
### 5. Community Jury (`community-jury.md`)
**"Randomly selected community members evaluate evidence and resolve disputes through structured deliberation"**
**Best for:**
- Formal dispute decisions
- Policy interpretation
- Agreement violations
- Cases needing peer judgment
**Key features:**
- Random selection (sortition)
- 5-7 member juries
- Structured evidence presentation
- Deliberation and voting
- Written decisions
- Creates precedent
**Complexity:** High | **Time:** 2-3 weeks | **Formality:** High
---
### 6. Community Referee (`community-referee.md`)
**"A single trained referee facilitates and decides on dispute resolution"**
**Best for:**
- Quick resolution needed
- Routine disputes
- Clear standards application
- Efficient use of resources
**Key features:**
- Single trained referee
- Streamlined process
- Expert decision-making
- Written determination
- Appeal to panel available
- Consistent application
**Complexity:** Medium | **Time:** 1-2 weeks | **Formality:** High
---
### 7. Facilitation Council (`facilitation-council.md`)
**"A trained council of facilitators manages the dispute resolution process"**
**Best for:**
- Complex cases needing multiple perspectives
- Balance of thoroughness and structure
- Communities with trained facilitators
- Cases affecting broader community
**Key features:**
- Panel of 2-3 facilitators
- Multi-session deliberation
- Tiered resolution approach
- Consensus prioritized
- Binding decisions when needed
- Comprehensive documentation
**Complexity:** High | **Time:** 2-4 weeks | **Formality:** High
---
### 8. Shalish Mediation (`shalish-mediation.md`)
**"Modernized traditional village-level mediation with cultural sensitivity"**
**Best for:**
- Communities with cultural traditions
- Harmony-focused resolution
- Relationship preservation
- Voluntary consensus-building
**Key features:**
- Traditional roots, modern adaptations
- Trained mediators
- Voluntary participation
- Cultural respect
- Community harmony focus
- No imposed solutions
**Complexity:** Medium | **Time:** Days to weeks | **Formality:** Low-Medium
---
## Comparison Matrix
| Template | Decision Authority | Speed | Formality | Community Involvement | Best For |
|----------|-------------------|-------|-----------|----------------------|----------|
| Peer-to-Peer | Parties | Fast | Low | Minimal | Minor conflicts |
| Chosen Facilitator | Parties | Medium | Medium | Low | Structured dialogue |
| Restorative Justice | Consensus | Slow | Medium | High | Healing harm |
| Transformative Justice | Parties + Community | Very Slow | Medium | Very High | Deep transformation |
| Community Jury | Jury | Medium | High | Medium | Formal decisions |
| Community Referee | Referee | Fast | High | Low | Efficient resolution |
| Facilitation Council | Council/Parties | Medium | High | Medium | Complex balanced cases |
| Shalish Mediation | Parties | Medium | Low-Medium | Medium | Cultural harmony |
## Choosing the Right Process
### Consider These Factors:
**Conflict Severity:**
- Minor → Peer-to-Peer
- Moderate → Chosen Facilitator, Referee
- Serious → Restorative Justice, Facilitation Council, Jury
- Systemic → Transformative Justice
**Decision Needs:**
- Parties decide → Peer-to-Peer, Chosen Facilitator, Shalish
- Binding decision needed → Referee, Jury, Council
- Consensus focus → Restorative Justice, Transformative Justice
**Time Available:**
- Quick (days) → Peer-to-Peer, Referee
- Moderate (weeks) → Most processes
- Extended (months) → Transformative Justice
**Community Resources:**
- Limited → Peer-to-Peer, Referee
- Moderate → Chosen Facilitator, Shalish
- Substantial → Circles, Jury, Council, Transformative Justice
**Cultural Context:**
- Western/modern → Most processes
- Traditional/cultural → Shalish, Circles
- Liberation-focused → Transformative Justice
## Implementation Recommendations
### Starting Out
If your community is new to formal dispute resolution:
1. Start with **Peer-to-Peer** for simple conflicts
2. Add **Chosen Facilitator** when trained facilitators available
3. Build toward more complex processes as capacity grows
### Comprehensive System
For mature communities, implement multiple pathways:
- **Peer-to-Peer** → First attempt
- **Chosen Facilitator or Shalish** → If peer-to-peer doesn't work
- **Restorative Circles** → For harm needing community involvement
- **Referee or Jury** → For formal binding decisions
- **Transformative Justice** → For systemic issues
### Essential Elements
Regardless of process chosen, include:
- Clear initiation procedures
- Ground rules for respectful engagement
- Voluntary participation (with consequences for refusal)
- Confidentiality protections
- Implementation and follow-up mechanisms
- Appeals or reconsideration pathways
- Learning and improvement systems
## Customization Tips
When adapting templates:
1. **Adjust to your scale** - Processes designed for 50-500 member communities may need modification for very small or very large groups
2. **Cultural adaptation** - Incorporate your community's values, traditions, and communication styles
3. **Resource reality** - Scale processes to available trained facilitators, time, and support
4. **Integration** - Ensure dispute resolution integrates smoothly with your governance constitution
5. **Bot capabilities** - Configure what the bot can automate (scheduling, reminders, documentation) vs. what requires human facilitation
6. **Language** - Keep natural language clear for bot interpretation while being specific enough for consistent application
## Training and Capacity Building
Most processes require trained facilitators:
- **Basic:** Peer-to-peer (minimal training)
- **Intermediate:** Chosen Facilitator, Shalish, Referee
- **Advanced:** Restorative Circles, Jury facilitation, Council, Transformative Justice
Consider:
- How will facilitators be trained?
- Who provides training?
- How are facilitators selected?
- Ongoing support and development?
- Community investment in capacity building
## Combining with Governance Templates
These dispute resolution templates work with any governance constitution:
- **Benevolent Dictator** → Add Peer-to-Peer and Referee for member conflicts
- **Do-ocracy** → Add Chosen Facilitator for when initiative creates conflict
- **Consensus** → Add Restorative or Transformative Justice (values alignment)
- **Circles** → Add Facilitation Council (similar structure)
- **Elected Board** → Add Referee or Jury (formal decisions)
- **Jury (governance)** → Add Community Jury for disputes too
- **Petition** → Add any process, chosen by petition
## Bot Implementation Notes
When configuring your governance bot:
1. **Trigger recognition** - Bot should recognize dispute resolution requests
2. **Process routing** - Direct to appropriate process based on situation
3. **Logistics support** - Scheduling, notifications, reminders
4. **Documentation** - Record keeping and confidentiality
5. **Tracking** - Monitor implementation and compliance
6. **Learning** - Collect anonymized data for improvement
7. **Escalation** - Enable movement between processes
## Credits
These templates are adapted from the [Dispute Protocol Builder](https://git.medlab.host/dispute-protocol/builder-prototype) project by the Media Enterprise Design Lab at the University of Colorado Boulder, with modifications for agentic governance bot integration.
## Further Resources
- **CommunityRule** - https://communityrule.info/ - Governance templates
- **Dispute Protocol Builder** - https://git.medlab.host/dispute-protocol/builder-prototype
- **Restorative Justice** - Various community resources on restorative circles
- **Transformative Justice** - Resources from community accountability movements
## Contributing
Found issues or have improvements? Contributions welcome to help these templates serve more communities effectively.
## License
[To be determined - match project license]