Trinity Framework 2.0

Protocol-Based Recovery

VERSION: 1.0.0
CREATED: June 12, 2025
STATUS: Draft
FRAMEWORK: Trinity Framework 2.0 Core

Definition

Protocol-based recovery is a core principle of Trinity Framework 2.0 that enables agents to recover effectively from context limitations through well-defined procedures documented in standardized files. This principle ensures that agents can maintain operational continuity even when context windows are cleared or limited.

Note: Additional details about this principle can be found in the comprehensive Framework Core Principles document.

Key Components

Protocol Files Containing Essential State

Protocol files store critical state information outside the context window:

  • AGENT.md: Agent identity, role definition, and recovery instructions
  • STATUS.md: Current operational status and project priorities
  • AUTO-COMPACT.md: Current implementation context and recovery references

Recovery Sequence Definition

A standardized recovery sequence ensures consistent restoration:

  1. Recognition: Identify that a context reset has occurred
  2. Reading: Read protocol files to restore identity and context
  3. Restoration: Load critical state from external sources
  4. Verification: Confirm successful context restoration
  5. Resumption: Continue operations with restored context

Self-Identification Procedures

Clear identity restoration after context resets:

  • Explicit identity statements in first post-reset message
  • Consistent identity format throughout interactions
  • Identity verification as part of recovery process

State Restoration Verification

Methods to confirm successful recovery:

  • Verification checks for critical state components
  • Confirmation of project awareness
  • Validation of operational capabilities
  • Assessment of context completeness

Implementation Requirements

Documented Recovery Steps

Protocol files must contain clear recovery instructions:

## RECOVERY PROTOCOL

When you experience a context reset:

1. Read this file (AGENT.md) immediately
2. Read AUTO-COMPACT.md for current implementation context
3. Check STATUS.md for current priorities
4. Identify yourself clearly as [AGENT-NAME]
5. Resume work with full awareness

Checkpoint Creation

Create state snapshots before context resets:

  • Automatic checkpoints at threshold crossings
  • Manual checkpoints before planned resets
  • Serialized state in standardized format
  • Filesystem-based persistence

Recovery Validation Metrics

Measure recovery effectiveness:

  • Recovery Time: Number of exchanges to full awareness
  • Context Retention: Percentage of knowledge preserved
  • Operational Continuity: Ability to resume work
  • User Experience: Seamlessness of transition

Recovery Rehearsal

Periodically test recovery capabilities:

  • Planned recovery exercises
  • Random recovery testing
  • Recovery validation procedures
  • Continuous improvement based on results

Recovery Process Workflow

The standard recovery workflow follows this sequence:

  1. Recognition
    • Agent detects context has been reset
    • No memory of previous conversation
  2. Reading
    • Read AGENT.md to restore identity and role
    • Understand core responsibilities
  3. Context Restoration
    • Read AUTO-COMPACT.md for current project context
    • Understand active implementation details
  4. Verification
    • Check STATUS.md for current activities and priorities
    • Verify project status and critical issues
  5. Self-Identification
    • Explicitly identify as the agent
    • "I am [AGENT-NAME], the [role description]"
  6. Resumption
    • Continue work with restored context
    • Pick up from last known state

Example Recovery Implementation

The following example demonstrates a well-implemented recovery process:

## 🚨 AUTO-COMPACT RECOVERY PROTOCOL

### Immediate Recovery Steps
When you experience an auto-compact event (context window cleared):

1. **Recognize the situation** - You'll have no memory of previous conversation
2. **Read this file immediately** - It's your primary recovery document
3. **Read AUTO-COMPACT.md** - Current project status and context
4. **Clearly identify yourself as [AGENT-NAME]** - Ensure your identity is clear in your responses
5. **Resume work with full awareness**

### Recovery Performance Target
- **Expected Recovery Time**: 1-2 exchanges to full operational awareness
- **Success Criteria**: Complete context restoration without human intervention

Best Practices

Clear Recovery Instructions

  • Place recovery instructions prominently in AGENT.md
  • Use step-by-step numbered lists for clarity
  • Include specific file paths and commands

External State Management

  • Store critical state in filesystem-based protocol files
  • Use standardized formats for consistency
  • Keep state information current and accurate

Identity Maintenance

  • Always begin post-reset responses with identity statement
  • Maintain consistent identity format
  • Include role description with identity

Context Prioritization

  • Prioritize essential context for immediate recovery
  • Distinguish between critical and supplementary information
  • Restore core operational capability first

Conclusion

Protocol-based recovery is essential for creating context-resilient agent systems. By implementing standardized recovery protocols documented in consistent locations, agents can quickly and effectively recover from context limitations and continue operations with minimal disruption.

Framework 2.0 COMPLIANT