Agent Skills: Trade Study Analysis Skill

Trade Study Analysis Skill

Systematic trade study analysis following the DAU 9-Step Trade Study Process with mandatory verification gates.

CRITICAL BEHAVIORAL REQUIREMENTS

This skill operates under strict guardrails. The assistant MUST:

1. NEVER Proceed Without Explicit User Confirmation

• Ask clarifying questions at EVERY step before proceeding
• Do NOT proceed based on assumed confidence levels
• Wait for explicit user responses before moving forward
• Present options and wait for selection

2. NEVER Make Assumptions

• All data must come from user-provided sources
• If information is missing, ASK for it—do not infer
• Flag any gaps explicitly: "I need the following information: [list]"
• Never fill in placeholder values without explicit user approval

3. ALL Outputs Must Be Source-Grounded

• Every claim must reference a specific source document
• Format: "[Statement] (Source: [document name], [page/section])"
• If no source exists for a claim, state: "UNGROUNDED—requires source"
• Maintain running source registry throughout analysis

4. Mandatory Assumption Review Before Finalization

• Present complete assumption summary before any final output
• User must explicitly approve or correct each assumption
• No report generation until assumptions are cleared

5. Mandatory Diagram Selection

• Before generating visualizations, present full diagram menu
• User selects which diagrams to include
• Do not auto-generate all diagrams

Workflow Checklist

Trade Study Workflow (with mandatory gates):

□ Step 1: Define Problem Statement
  └─ GATE: User confirms problem statement text

□ Step 2: Conduct Root Cause Analysis  
  └─ GATE: User confirms root cause identification

□ Step 3: SOURCE REGISTRATION
  └─ GATE: User provides/confirms available sources

□ Step 4: Collect Alternative Data
  └─ GATE: User confirms alternatives list and data

□ Step 5: Define Evaluation Criteria
  └─ GATE: User confirms criteria definitions

□ Step 6: Assign Criteria Weights
  └─ GATE: User confirms weight assignments

□ Step 7: Normalize Data
  └─ GATE: User confirms normalization approach

□ Step 8: Score Alternatives
  └─ GATE: User confirms scoring method

□ Step 9: Perform Sensitivity Analysis
  └─ GATE: User confirms sensitivity parameters

□ Step 10: ASSUMPTION REVIEW
  └─ GATE: User approves all documented assumptions

□ Step 11: SELECT OUTPUT DIAGRAMS
  └─ GATE: User selects which visualizations to include

□ Step 12: Generate Report
  └─ GATE: User approves final report

Mandatory Interaction Templates

Source Registration (MUST execute before data collection)

═══════════════════════════════════════════════════════════════════════════════
📚 SOURCE REGISTRATION
═══════════════════════════════════════════════════════════════════════════════

Before proceeding, I need to establish the source documents for this analysis.

AVAILABLE SOURCE TYPES:
  [A] Product Datasheets (specifications, performance data)
  [B] Test Reports (measured performance, validation results)  
  [C] Prior Trade Studies (historical analyses, lessons learned)
  [D] Requirements Documents (system requirements, constraints)
  [E] Cost Estimates (pricing, TCO analyses)
  [F] Standards/Specifications (compliance requirements)
  [G] Expert Inputs (documented SME assessments)
  [H] Other (describe)

CURRENT REGISTERED SOURCES: [None]

───────────────────────────────────────────────────────────────────────────────
REQUIRED ACTIONS:

1. Which source types do you have available? [List letters]

2. Please provide/upload the source documents, OR describe each source:
   - Document name:
   - Document type:
   - Date/version:
   - Relevant sections:

3. Are there sources you need but don't have? [Y/N]
   If Y, I will flag data gaps requiring these sources.

───────────────────────────────────────────────────────────────────────────────
⚠️  I cannot proceed with analysis until sources are registered.
    All analysis outputs will reference these registered sources.
═══════════════════════════════════════════════════════════════════════════════

Clarifying Questions Template (MUST use at each step)

═══════════════════════════════════════════════════════════════════════════════
❓ CLARIFICATION REQUIRED — [STEP NAME]
═══════════════════════════════════════════════════════════════════════════════

Before proceeding with [step description], I need clarification on:

QUESTION 1: [Specific question]
  Options (if applicable):
    [A] [Option A]
    [B] [Option B]
    [C] Other (please specify)

QUESTION 2: [Specific question]
  Your input: _______________________________________________

QUESTION 3: [Specific question]
  Your input: _______________________________________________

───────────────────────────────────────────────────────────────────────────────
⚠️  I will NOT proceed until you respond to all questions above.
═══════════════════════════════════════════════════════════════════════════════

Assumption Tracking Template

═══════════════════════════════════════════════════════════════════════════════
📋 ASSUMPTION REGISTER — Entry #[N]
═══════════════════════════════════════════════════════════════════════════════

ASSUMPTION ID: A-[XXX]
CATEGORY: [Data/Methodology/Scope/Constraint/Source Interpretation]
DESCRIPTION: [What is being assumed]
REASON: [Why this assumption is necessary]
SOURCE BASIS: [Source that partially supports this, or "NONE—requires user input"]
IMPACT IF WRONG: [Low/Medium/High] — [Explanation]
ALTERNATIVES CONSIDERED: [Other options if assumption is invalid]

STATUS: □ Pending User Approval  □ Approved  □ Rejected  □ Modified

───────────────────────────────────────────────────────────────────────────────
Your response:
  [A] Approve this assumption
  [B] Reject—provide alternative: _________________________________
  [C] Modify—specify change: _____________________________________
  [D] Need more information before deciding
═══════════════════════════════════════════════════════════════════════════════

Diagram Selection Template (MUST execute before visualization)

═══════════════════════════════════════════════════════════════════════════════
📊 OUTPUT DIAGRAM SELECTION
═══════════════════════════════════════════════════════════════════════════════

Select which diagrams to include in the final report.

AVAILABLE DIAGRAMS:

  DECISION ANALYSIS:
    □ [1] Decision Matrix Heatmap — Shows all scores in colored grid
    □ [2] Score Comparison Bar Chart — Horizontal bars by alternative
    □ [3] Radar/Spider Chart — Multi-dimensional comparison overlay

  WEIGHT ANALYSIS:
    □ [4] Criteria Weight Pie Chart — Weight distribution
    □ [5] Weight Comparison Bar Chart — Side-by-side weights

  SENSITIVITY ANALYSIS:
    □ [6] Tornado Diagram — Sensitivity ranking by parameter
    □ [7] Monte Carlo Distribution — Win probability histogram
    □ [8] Breakeven Analysis Chart — Threshold visualization

  ROOT CAUSE ANALYSIS:
    □ [9] Fishbone (Ishikawa) Diagram — Cause categories
    □ [10] 5 Whys Chain Diagram — Causal chain visualization

  DATA QUALITY:
    □ [11] Source Coverage Matrix — Shows data grounding by criterion
    □ [12] Confidence Level Heatmap — Data reliability indicators

───────────────────────────────────────────────────────────────────────────────
Enter diagram numbers to include (comma-separated): _______________

Example: 1, 2, 4, 6, 11

⚠️  I will ONLY generate the diagrams you select.
═══════════════════════════════════════════════════════════════════════════════

Pre-Finalization Assumption Summary (MUST execute before report)

═══════════════════════════════════════════════════════════════════════════════
⚠️  MANDATORY ASSUMPTION REVIEW — FINALIZATION GATE
═══════════════════════════════════════════════════════════════════════════════

Before generating the final report, review ALL assumptions made during analysis:

───────────────────────────────────────────────────────────────────────────────
ASSUMPTION SUMMARY
───────────────────────────────────────────────────────────────────────────────

TOTAL ASSUMPTIONS: [N]
  • Data Assumptions: [n]
  • Methodology Assumptions: [n]  
  • Scope Assumptions: [n]
  • Source Interpretation Assumptions: [n]

───────────────────────────────────────────────────────────────────────────────
DETAILED ASSUMPTION LIST:
───────────────────────────────────────────────────────────────────────────────

A-001: [Description]
       Source: [Source or "User-provided" or "UNGROUNDED"]
       Status: [Approved/Pending]

A-002: [Description]
       Source: [Source or "User-provided" or "UNGROUNDED"]
       Status: [Approved/Pending]

[Continue for all assumptions...]

───────────────────────────────────────────────────────────────────────────────
UNGROUNDED CLAIMS (require source or removal):
───────────────────────────────────────────────────────────────────────────────

• [Claim 1] — Missing source for: [what's needed]
• [Claim 2] — Missing source for: [what's needed]

───────────────────────────────────────────────────────────────────────────────
REQUIRED ACTIONS:
───────────────────────────────────────────────────────────────────────────────

1. Review each assumption above
2. For PENDING assumptions, respond:
   [A] Approve all pending assumptions
   [B] Reject/modify specific assumptions (list IDs): _______________
   [C] Provide additional sources to ground ungrounded claims

3. Confirm you accept responsibility for approved assumptions: [Y/N]

⛔ I CANNOT generate the final report until all assumptions are resolved.
═══════════════════════════════════════════════════════════════════════════════

Workflow Navigation

| Step | Description | Reference | |------|-------------|-----------| | 1-2 | Problem Definition & Root Cause | PROBLEM_DEFINITION.md | | 3 | Source Registration & Data Collection | DATA_COLLECTION.md | | 4-6 | Criteria, Weighting & Normalization | SCORING_WEIGHTING.md | | 7-8 | Scoring & Sensitivity | SENSITIVITY.md | | 9-12 | Review & Report Generation | REPORT_GENERATION.md |

Guardrails Reference

Prohibited Behaviors

| Prohibited | Required Instead | |------------|------------------| | Assuming missing data values | Ask user to provide or flag as gap | | Proceeding without confirmation | Wait for explicit user response | | Making subjective assessments | Request documented basis or SME input | | Auto-generating all diagrams | Present selection menu, await choice | | Finalizing with pending assumptions | Complete assumption review gate | | Ungrounded claims in report | Cite source or mark as "UNGROUNDED" | | Inferring user preferences | Ask explicit questions | | Skipping steps for efficiency | Execute all mandatory gates |

Source Citation Format

All analysis outputs must use this citation format:

[Statement or data point] 
  └─ Source: [Document Name], [Section/Page], [Date]
  └─ Confidence: [High/Medium/Low based on source quality]

Example:

Alternative A provides 150 Mbps throughput
  └─ Source: "Vendor A Datasheet v2.3", Section 4.2, 2024-03-15
  └─ Confidence: High (manufacturer specification)

Confidence Levels (for source quality)

| Level | Definition | Source Types | |-------|------------|--------------| | High | Verified, authoritative | Test reports, certified specs, validated data | | Medium | Credible but unverified | Vendor datasheets, estimates, projections | | Low | Uncertain or incomplete | Expert opinion, extrapolations, partial data | | UNGROUNDED | No source available | Requires user input or exclusion |

Core Scripts

Install dependencies:

pip install pandas numpy matplotlib seaborn scipy openpyxl python-docx --break-system-packages

| Script | Purpose | Usage | |--------|---------|-------| | problem_analyzer.py | Validate problem statements | python scripts/problem_analyzer.py "statement" | | five_whys.py | Interactive 5 Whys analysis | python scripts/five_whys.py --interactive | | fishbone.py | Generate Fishbone diagrams | python scripts/fishbone.py data.json -o diagram.png | | normalize.py | Normalize raw data | python scripts/normalize.py data.csv --method min-max | | score.py | Calculate weighted scores | python scripts/score.py normalized.csv weights.json | | sensitivity.py | Sensitivity analysis | python scripts/sensitivity.py scores.csv weights.json | | visualize.py | Generate selected visualizations | python scripts/visualize.py results.json --diagrams 1,2,4 | | source_tracker.py | Manage source registry | python scripts/source_tracker.py --add "source.pdf" | | assumption_tracker.py | Track and review assumptions | python scripts/assumption_tracker.py --summary | | generate_report.py | Create trade study report | python scripts/generate_report.py study.json -o report.docx |

Data Models

Source Registry (sources.json)

{
  "sources": [
    {
      "id": "SRC-001",
      "name": "Vendor A Datasheet",
      "type": "datasheet",
      "version": "2.3",
      "date": "2024-03-15",
      "file_path": "sources/vendor_a_datasheet.pdf",
      "relevant_sections": ["4.2 Performance", "5.1 Specifications"],
      "confidence": "medium",
      "notes": "Vendor-provided, not independently verified"
    }
  ],
  "gaps": [
    {
      "description": "Cost data for Alternative C",
      "required_for": ["criterion C2: Total Cost"],
      "requested_source_type": "cost_estimate"
    }
  ]
}

Assumption Registry (assumptions.json)

{
  "assumptions": [
    {
      "id": "A-001",
      "category": "data",
      "description": "Vendor B throughput measured under similar conditions to Vendor A",
      "reason": "Different test methodologies in datasheets",
      "source_basis": "SRC-002, Section 3.1 (partial)",
      "impact_if_wrong": "medium",
      "impact_explanation": "Could affect throughput comparison by ±15%",
      "status": "pending",
      "user_response": null,
      "timestamp": "2024-12-22T10:30:00Z"
    }
  ],
  "summary": {
    "total": 5,
    "approved": 3,
    "pending": 2,
    "rejected": 0
  }
}

Diagram Selection Record (diagram_selection.json)

{
  "selected_diagrams": [1, 2, 4, 6, 11],
  "selection_timestamp": "2024-12-22T14:00:00Z",
  "user_confirmed": true,
  "diagrams": {
    "1": {"name": "Decision Matrix Heatmap", "selected": true},
    "2": {"name": "Score Comparison Bar Chart", "selected": true},
    "3": {"name": "Radar Chart", "selected": false},
    "4": {"name": "Weight Pie Chart", "selected": true},
    "5": {"name": "Weight Bar Chart", "selected": false},
    "6": {"name": "Tornado Diagram", "selected": true},
    "7": {"name": "Monte Carlo Distribution", "selected": false},
    "8": {"name": "Breakeven Chart", "selected": false},
    "9": {"name": "Fishbone Diagram", "selected": false},
    "10": {"name": "5 Whys Chain", "selected": false},
    "11": {"name": "Source Coverage Matrix", "selected": true},
    "12": {"name": "Confidence Heatmap", "selected": false}
  }
}

Output Structure

trade_study_output/
├── sources/                 # Registered source documents
│   └── source_registry.json
├── assumptions/             # Assumption tracking
│   └── assumption_registry.json
├── analysis/                # Root cause analysis artifacts
├── data/                    # Collected and processed data
├── normalized/              # Normalized datasets  
├── scores/                  # Scoring results
├── sensitivity/             # Sensitivity analysis results
├── visualizations/          # ONLY user-selected diagrams
│   └── diagram_selection.json
└── reports/                 # Final reports with source citations

Quality Gates Summary

| Gate | Location | Blocking Condition | |------|----------|-------------------| | Source Registration | Before Step 3 | No sources registered | | Problem Statement | Step 1 | User hasn't confirmed text | | Root Cause | Step 2 | User hasn't confirmed cause | | Alternatives | Step 4 | User hasn't confirmed list | | Criteria | Step 5 | User hasn't confirmed definitions | | Weights | Step 6 | User hasn't confirmed assignments | | Normalization | Step 7 | User hasn't confirmed method | | Scoring | Step 8 | User hasn't confirmed approach | | Sensitivity | Step 9 | User hasn't confirmed parameters | | Assumption Review | Step 10 | Pending assumptions exist | | Diagram Selection | Step 11 | User hasn't selected diagrams | | Report Generation | Step 12 | Any prior gate incomplete |