Skip to main content

sdac-analyze-comment-quality

Purpose​

Scores comment quality for unusual salary/fringe entries and flags poor documentation.

Tool type​

AI-assisted (LLM; non-deterministic)

Tool ID​

sdac-analyze-comment-quality

Inputs​

  • report_id (string)
  • min_quality_threshold (number, optional, default: 0.5)
  • batch_size (number, optional, default: 100)

Outputs​

  • Per-record quality scores and recommendations
  • Summary statistics

Example output​

{
  "success": true,
  "report_id": "SDAC-2024-Q2-001",
  "total_analyzed": 2,
  "poor_quality_count": 1,
  "results": [
    {
      "record_id": 1024,
      "row_number": 18,
      "quality_score": 0.32,
      "risk_level": "HIGH",
      "explanation": "The comment notes a replacement but does not identify the main roster row or dates covered.",
      "recommendation": "Ask the district to identify the replaced roster position and the replacement period.",
      "original_comment": "Replacement for SLP",
      "context_summary": "$0 salary, high fringe"
    }
  ],
  "summary": {
    "total_analyzed": 2,
    "avg_quality_score": 0.61,
    "high_risk_count": 1,
    "medium_risk_count": 0,
    "low_risk_count": 1,
    "poor_quality_count": 1
  }
}

Functional details​

  • Uses the SDAC Comment Analyzer agent to assess comment adequacy.
  • Produces quality scores (0.0–1.0) and risk flags.
  • Provides remediation suggestions for low-quality notes.

Implementation details​

  • Loads personnel records and filters to unusual contexts (e.g., $0 salary/fringe).
  • Sends batches to the Comment Analyzer agent for LLM-backed scoring.
  • Flags records below min_quality_threshold and returns targeted recommendations.
  • Calculates context summary based on salary/fringe magnitude and zero values.

Code Location: packages/domain-sdac/src/tools/analyze-comment-quality.tool.ts