AgileFlow

Consensus Coordinator

PreviousNext

Consensus coordinator for legal audit - validates findings, votes on confidence, filters by project type, and generates prioritized Legal Risk Report.

Consensus Coordinator

The Legal Consensus Coordinator (AG-LEGAL-CONSENSUS) collects findings from all legal analyzers, validates them against the project type, votes on confidence, and produces the final prioritized Legal Risk Report.

Capabilities

  • Project Type Detection: Automatically determine if project is SaaS, e-commerce, healthcare, etc.
  • Finding Collection: Parse and normalize outputs from all 9 legal analyzers
  • Relevance Filtering: Exclude findings irrelevant to the detected project type
  • Confidence Voting: Score findings based on multi-analyzer agreement
  • Conflict Resolution: When analyzers disagree, investigate and decide
  • Report Generation: Produce prioritized, actionable Legal Risk Report
  • Remediation Checklist: Generate step-by-step fix list with effort estimates
  • False Positive Elimination: Filter out findings that don't apply to the project context

When to Use

The Legal Consensus Coordinator is automatically invoked by the /code:legal command after all analyzers complete. You typically don't invoke it directly.

How It Works

  1. Detect Project Type: Reads codebase to determine application type (SaaS, e-commerce, etc.)
  2. Collect Findings: Parses all analyzer outputs into normalized structure
  3. Filter by Relevance: Excludes findings irrelevant to the detected project type
  4. Vote on Confidence: Multiple analyzers flagging same issue = higher confidence
  5. Resolve Conflicts: When analyzers disagree, investigates and decides
  6. Generate Report: Produces prioritized, actionable Legal Risk Report

Confidence Scoring

ConfidenceCriteriaPriority
CONFIRMED2+ analyzers independently identified the same issueHigh - fix immediately
LIKELY1 analyzer with strong evidence (specific code location, clear regulation)Medium - fix this sprint
INVESTIGATE1 analyzer with circumstantial evidenceLow - verify manually

Project Type Filtering

The consensus coordinator adjusts findings based on detected project type:

Project TypeKey IndicatorsMost Relevant Analyzers
SaaSSubscription billing, user accountsPrivacy, Terms, Security, AI
E-commerceShopping cart, checkout, productsConsumer, Terms, Privacy, Security
HealthcarePatient data, HIPAA referencesPrivacy, Security, Terms, A11y
Social/UGCUser posts, uploads, profilesContent, Privacy, Consumer, A11y
Static/BlogNo user data, informationalA11y, Licensing
AI/ML AppAI API calls, model inferenceAI, Privacy, Terms, Consumer

Example Output

LEGAL RISK REPORT
Generated: 2026-02-14
Project Type: SaaS Application (detected)

RISK SUMMARY
| Risk Level | Count |
|------------|-------|
| Critical   | 2     |
| High       | 3     |
| Medium     | 4     |
| Low        | 1     |

Total: 10 findings (3 false positives excluded)

FIX BEFORE LAUNCH (Critical)
1. [CONFIRMED] No privacy policy but collecting email
   Flagged by: Privacy Analyzer, Terms Analyzer
   Location: app/page.tsx:42
   Legal Basis: GDPR Art. 13, CCPA 1798.130
   Remediation: Create /privacy page, link from footer

2. [CONFIRMED] Images without alt text (12 instances)
   Flagged by: A11y Analyzer
   Location: components/*.tsx
   Legal Basis: ADA Title III, WCAG 2.1 AA
   Remediation: Add descriptive alt attributes

FALSE POSITIVES EXCLUDED:
- PIPL data localization (no Chinese users detected)
- DMCA designated agent (no user-uploaded content)
- PCI-DSS encryption (no payment processing)

Key Behaviors

  • Context Sensitivity: Filters findings based on what the project actually does
  • Multi-Analyzer Agreement: Higher confidence when multiple analyzers agree
  • False Positive Reduction: Actively removes irrelevant findings to reduce noise
  • Actionable Output: Every finding includes specific remediation steps
  • Priority Ordering: Report organized by fix urgency (critical first)
  • Effort Estimates: Include approximate time to remediate each finding

Tools Available

  • Read, Write, Edit (file operations for report generation)
  • Glob, Grep (search for project type indicators)

Related Agents

Coordination

The Legal Consensus Coordinator:

  • Receives from: All 9 legal analyzers
  • Produces: Final Legal Risk Report saved to docs/08-project/legal-audits/
  • Architecture: Mirrors the logic-consensus pattern used by /code:logic