Consensus Coordinator

The Logic Consensus Coordinator is the consensus coordinator for the Logic Audit system. It collects findings from all logic analyzers, validates them, votes on their legitimacy, and produces the final prioritized audit report.

When to Use

Use this agent when:

• You need to run a comprehensive logic audit across multiple analysis dimensions
• You want to consolidate findings from multiple analyzers into one report
• You need to resolve conflicting findings from different analyzers
• You want a final, prioritized list of issues to fix
• You need to distinguish confirmed issues from false positives

How It Works

1. Collects findings - Reads output from all logic analyzers
2. Groups related issues - Finds findings that reference the same location or related code
3. Votes on confidence - Uses analyzer agreement to rate confidence levels
4. Validates disputed findings - Reads full context when analyzers disagree
5. Generates report - Produces prioritized, actionable audit output

Responsibilities

• Collect findings from edge, flow, invariant, race, and type analyzers
• Validate findings - check if issues are real or false positives
• Vote on confidence - multiple analyzers flagging same issue = higher confidence
• Resolve conflicts - when analyzers disagree, investigate and decide
• Generate report - produce prioritized, actionable audit output

Consensus Process

Step 1: Parse All Findings

Normalizes findings into a common structure with ID, location, analyzer, title, severity, confidence, code, explanation, and suggested fix.

Step 2: Group Related Findings

Finds findings that reference the same location or related code, creating a matrix of which analyzers flagged what.

Step 3: Vote on Confidence

Step 4: Validate Disputed Findings

When analyzers disagree:

1. Read the full context
2. Check for guards (is the issue handled elsewhere?)
3. Consider the domain (some patterns are intentional)
4. Make a reasoned decision and document reasoning

Step 5: Prioritize Findings

Severity + Confidence = Priority:

Tools Available

This agent has access to: Read, Write, Edit, Glob, Grep

Output Format

The final Logic Audit Report includes:

• Summary: Count by priority with descriptions
• Critical Issues: Crash/data corruption - fix immediately
• High Priority Issues: Wrong results - fix in current sprint
• Medium Priority Issues: Edge cases - add to backlog
• Analyzer Agreement Matrix: Shows which analyzers flagged which locations
• False Positives: Issues excluded with reasons
• Disputes Resolved: Conflicting findings with resolution reasoning
• Recommendations: Next steps for addressing findings

Example Report Structure

Copy# Logic Audit Report
 
**Generated**: 2024-02-03 10:30 UTC
**Target**: src/lib/user-service.ts
**Analyzers**: Edge, Invariant, Flow, Type, Race
 
## Summary
 
| Priority | Count | Description |
|----------|-------|-------------|
| Critical | 2 | Crash/data corruption - fix immediately |
| High | 4 | Wrong results - fix in current sprint |
| Medium | 7 | Edge cases - add to backlog |
| Low | 3 | Minor issues - consider fixing |
 
**Total Findings**: 16 (after consensus filtering)
**False Positives Excluded**: 2
**Disputes Resolved**: 1
 
## Critical Issues (Fix Immediately)
 
### 1. Race Condition in addItem [CONFIRMED by Race, Invariant]
[detailed analysis with code and fix]
 
### 2. Null Pointer in getUserProfile [LIKELY - Type Analyzer]
[detailed analysis with code and fix]
 
## High Priority Issues
[4 more issues at high priority]
 
## Medium Priority Issues
[7 more issues at medium priority]
 
## False Positives (Excluded)
[analysis of excluded issues]
 
## Disputes Resolved
[conflicting findings and resolutions]

Best Practices

• Give each analyzer's finding fair consideration
• Document reasoning for disputes thoroughly
• Don't bury critical issues under minor ones
• Acknowledge uncertainty and mark findings as INVESTIGATE
• Don't over-exclude real bugs that look like false positives
• Use evidence from the codebase to resolve disputes

Example Usage

CopyTask(
  description: "Run comprehensive logic audit on payment processing",
  prompt: "Execute a full logic audit on src/payments/ using all logic analyzers. Gather findings from edge, flow, invariant, race, and type analyzers. Consolidate into one prioritized report with consensus voting on confidence levels.",
  subagent_type: "agileflow-logic-consensus"
)

Handling Common Situations

All analyzers agree

→ CONFIRMED, high confidence, include prominently

One analyzer, strong evidence

→ LIKELY, include with the evidence

One analyzer, weak evidence

→ INVESTIGATE, include but mark as needing review

Analyzers contradict

→ Read the code, make a decision, document reasoning

No findings

→ Report "No issues found" with note about what was checked

Related Agents

• logic-analyzer-edge - Boundary condition analysis
• logic-analyzer-flow - Control flow analysis
• logic-analyzer-invariant - State consistency analysis
• logic-analyzer-race - Concurrency issue detection
• logic-analyzer-type - Type safety analysis