plugin-agent-orchestrator 1.0.6 → 1.0.14

package/dist/server/services/SkillManager.js

@@ -0,0 +1,640 @@
+/**
+ * This file is part of the NocoBase (R) project.
+ * Copyright (c) 2020-2024 NocoBase Co., Ltd.
+ * Authors: NocoBase Team.
+ *
+ * This project is dual-licensed under AGPL-3.0 and NocoBase Commercial License.
+ * For more information, please refer to: https://www.nocobase.com/agreement.
+ */
+
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var SkillManager_exports = {};
+__export(SkillManager_exports, {
+  SkillManager: () => SkillManager
+});
+module.exports = __toCommonJS(SkillManager_exports);
+var import_json_fields = require("../skill-hub/utils/json-fields");
+class SkillManager {
+  constructor(db) {
+    this.db = db;
+  }
+  async seedDefaults() {
+    const repo = this.db.getRepository("skillDefinitions");
+    const seeds = [
+      {
+        name: "generate-word-report",
+        title: "Generate Word Report",
+        description: "Generate a Word document (.docx) with title, content, and optional table data.",
+        language: "python",
+        codeTemplate: SEED_WORD_REPORT,
+        inputSchema: {
+          type: "object",
+          properties: {
+            title: { type: "string", description: "Report title" },
+            content: { type: "string", description: "Report body text" },
+            tableData: {
+              type: "array",
+              items: { type: "object" },
+              description: "Optional array of objects for table rows"
+            }
+          },
+          required: ["title", "content"]
+        },
+        packages: ["python-docx"],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "generate-excel",
+        title: "Generate Excel Spreadsheet",
+        description: "Generate an Excel file (.xlsx) with headers and row data.",
+        language: "python",
+        codeTemplate: SEED_EXCEL,
+        inputSchema: {
+          type: "object",
+          properties: {
+            sheetName: { type: "string", description: "Sheet name", default: "Sheet1" },
+            headers: { type: "array", items: { type: "string" }, description: "Column headers" },
+            rows: { type: "array", items: { type: "array" }, description: "Row data (array of arrays)" }
+          },
+          required: ["headers", "rows"]
+        },
+        packages: ["openpyxl"],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "generate-pdf-report",
+        title: "Generate PDF Report",
+        description: "Generate a professional PDF report with title, sections, and optional table. Suitable for CRM reports, customer summaries, and sales reports.",
+        language: "python",
+        codeTemplate: SEED_PDF_REPORT,
+        inputSchema: {
+          type: "object",
+          properties: {
+            title: { type: "string", description: "Report title" },
+            sections: {
+              type: "array",
+              items: {
+                type: "object",
+                properties: {
+                  heading: { type: "string" },
+                  body: { type: "string" }
+                }
+              },
+              description: "Array of {heading, body} sections"
+            },
+            tableData: {
+              type: "array",
+              items: { type: "object" },
+              description: "Optional array of objects for a summary table"
+            }
+          },
+          required: ["title", "sections"]
+        },
+        packages: ["reportlab"],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "generate-chart-image",
+        title: "Generate Chart Image",
+        description: "Generate a chart image (PNG) from data. Supports bar, line, pie charts. Useful for visualizing CRM metrics like sales, leads, revenue.",
+        language: "python",
+        codeTemplate: SEED_CHART,
+        inputSchema: {
+          type: "object",
+          properties: {
+            chartType: {
+              type: "string",
+              enum: ["bar", "line", "pie"],
+              description: "Chart type"
+            },
+            title: { type: "string", description: "Chart title" },
+            labels: { type: "array", items: { type: "string" }, description: "X-axis labels or pie labels" },
+            values: { type: "array", items: { type: "number" }, description: "Data values" },
+            xlabel: { type: "string", description: "X-axis label (bar/line only)" },
+            ylabel: { type: "string", description: "Y-axis label (bar/line only)" }
+          },
+          required: ["chartType", "title", "labels", "values"]
+        },
+        packages: ["matplotlib"],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "generate-invoice-pdf",
+        title: "Generate Invoice PDF",
+        description: "Generate a professional invoice PDF with company info, line items, and totals. For CRM billing and quotation workflows.",
+        language: "python",
+        codeTemplate: SEED_INVOICE,
+        inputSchema: {
+          type: "object",
+          properties: {
+            invoiceNumber: { type: "string", description: "Invoice number" },
+            date: { type: "string", description: "Invoice date (YYYY-MM-DD)" },
+            companyName: { type: "string", description: "Your company name" },
+            customerName: { type: "string", description: "Customer name" },
+            customerAddress: { type: "string", description: "Customer address" },
+            items: {
+              type: "array",
+              items: {
+                type: "object",
+                properties: {
+                  description: { type: "string" },
+                  quantity: { type: "number" },
+                  unitPrice: { type: "number" }
+                }
+              },
+              description: "Line items with description, quantity, unitPrice"
+            },
+            currency: { type: "string", description: "Currency symbol", default: "$" },
+            notes: { type: "string", description: "Optional notes or payment terms" }
+          },
+          required: ["invoiceNumber", "date", "companyName", "customerName", "items"]
+        },
+        packages: ["reportlab"],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "data-summary-report",
+        title: "Data Summary & Analysis Report",
+        description: "Analyze tabular data and generate a Word report with summary statistics, top records, and insights. Ideal for CRM data analysis \u2014 leads, deals, customers.",
+        language: "python",
+        codeTemplate: SEED_DATA_SUMMARY,
+        inputSchema: {
+          type: "object",
+          properties: {
+            title: { type: "string", description: "Report title" },
+            data: {
+              type: "array",
+              items: { type: "object" },
+              description: "Array of data objects (e.g. deals, leads, customers)"
+            },
+            groupByField: { type: "string", description: "Optional field to group and summarize by" },
+            sortByField: { type: "string", description: "Optional field to sort by (descending)" },
+            topN: { type: "number", description: "Number of top records to include", default: 10 }
+          },
+          required: ["title", "data"]
+        },
+        packages: ["pandas", "python-docx"],
+        timeoutSeconds: 60,
+        enabled: true,
+        toolScope: "CUSTOM"
+      },
+      {
+        name: "data-transform",
+        title: "Data Transform (CSV/JSON)",
+        description: "Transform data array to CSV or JSON file for download.",
+        language: "node",
+        codeTemplate: SEED_DATA_TRANSFORM,
+        inputSchema: {
+          type: "object",
+          properties: {
+            data: { type: "array", items: { type: "object" }, description: "Array of data objects" },
+            format: { type: "string", enum: ["csv", "json"], description: "Output format" },
+            filename: { type: "string", description: "Output filename (without extension)" }
+          },
+          required: ["data", "format"]
+        },
+        packages: [],
+        timeoutSeconds: 30,
+        enabled: true,
+        toolScope: "CUSTOM"
+      }
+    ];
+    for (const seed of seeds) {
+      try {
+        const count = await repo.count({ filter: { name: seed.name } });
+        if (count === 0) {
+          await repo.create({
+            values: {
+              ...seed,
+              inputSchema: (0, import_json_fields.stringifyJsonText)(seed.inputSchema),
+              packages: (0, import_json_fields.stringifyJsonText)(seed.packages, [])
+            }
+          });
+        }
+      } catch (err) {
+        console.error(`[import-skill] Failed to insert ${seed.name}:`, err);
+      }
+    }
+  }
+}
+const SEED_WORD_REPORT = `import os, json
+from docx import Document
+
+title = json.loads('''{{title}}''') if '{{title}}'.startswith('"') else '{{title}}'
+content = json.loads('''{{content}}''') if '{{content}}'.startswith('"') else '{{content}}'
+
+doc = Document()
+doc.add_heading(title, 0)
+doc.add_paragraph(content)
+
+table_data_raw = '''{{tableData}}'''
+if table_data_raw and table_data_raw != '{{' + 'tableData}}':
+    table_data = json.loads(table_data_raw)
+    if table_data and len(table_data) > 0:
+        headers = list(table_data[0].keys())
+        table = doc.add_table(rows=1, cols=len(headers))
+        table.style = 'Light Grid Accent 1'
+        for i, header in enumerate(headers):
+            table.rows[0].cells[i].text = str(header)
+        for row_data in table_data:
+            row = table.add_row()
+            for i, header in enumerate(headers):
+                row.cells[i].text = str(row_data.get(header, ''))
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'report.docx')
+doc.save(filepath)
+print(f'Generated: report.docx')
+`;
+const SEED_EXCEL = `import os, json
+import openpyxl
+
+sheet_name_raw = '{{sheetName}}'
+sheet_name = sheet_name_raw if sheet_name_raw != '{{' + 'sheetName}}' else 'Sheet1'
+headers = json.loads('''{{headers}}''')
+rows = json.loads('''{{rows}}''')
+
+wb = openpyxl.Workbook()
+ws = wb.active
+ws.title = sheet_name
+
+for col, header in enumerate(headers, 1):
+    ws.cell(row=1, column=col, value=header)
+
+for row_idx, row_data in enumerate(rows, 2):
+    for col_idx, value in enumerate(row_data, 1):
+        ws.cell(row=row_idx, column=col_idx, value=value)
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'data.xlsx')
+wb.save(filepath)
+print(f'Generated: data.xlsx')
+`;
+const SEED_PDF_REPORT = `import os, json
+from reportlab.lib.pagesizes import A4
+from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
+from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, Table, TableStyle
+from reportlab.lib import colors
+from reportlab.lib.units import cm
+
+title = json.loads('''{{title}}''') if '{{title}}'.startswith('"') else '{{title}}'
+sections = json.loads('''{{sections}}''')
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'report.pdf')
+
+doc = SimpleDocTemplate(filepath, pagesize=A4,
+    topMargin=2*cm, bottomMargin=2*cm, leftMargin=2*cm, rightMargin=2*cm)
+styles = getSampleStyleSheet()
+title_style = ParagraphStyle('CustomTitle', parent=styles['Title'], fontSize=22, spaceAfter=20)
+heading_style = ParagraphStyle('CustomHeading', parent=styles['Heading2'], fontSize=14, spaceBefore=16, spaceAfter=8)
+body_style = styles['BodyText']
+
+story = [Paragraph(title, title_style), Spacer(1, 12)]
+
+for section in sections:
+    if section.get('heading'):
+        story.append(Paragraph(section['heading'], heading_style))
+    if section.get('body'):
+        for line in section['body'].split('\\n'):
+            story.append(Paragraph(line, body_style))
+    story.append(Spacer(1, 8))
+
+table_data_raw = '''{{tableData}}'''
+if table_data_raw and table_data_raw != '{{' + 'tableData}}':
+    td = json.loads(table_data_raw)
+    if td and len(td) > 0:
+        headers = list(td[0].keys())
+        data = [headers] + [[str(r.get(h, '')) for h in headers] for r in td]
+        t = Table(data, repeatRows=1)
+        t.setStyle(TableStyle([
+            ('BACKGROUND', (0, 0), (-1, 0), colors.HexColor('#4472C4')),
+            ('TEXTCOLOR', (0, 0), (-1, 0), colors.white),
+            ('FONTSIZE', (0, 0), (-1, 0), 10),
+            ('FONTSIZE', (0, 1), (-1, -1), 9),
+            ('GRID', (0, 0), (-1, -1), 0.5, colors.grey),
+            ('ROWBACKGROUNDS', (0, 1), (-1, -1), [colors.white, colors.HexColor('#D9E2F3')]),
+        ]))
+        story.append(Spacer(1, 12))
+        story.append(t)
+
+doc.build(story)
+print('Generated: report.pdf')
+`;
+const SEED_PPTX = `import os, json
+from pptx import Presentation
+from pptx.util import Inches, Pt
+from pptx.enum.text import PP_ALIGN
+
+title = json.loads('''{{title}}''') if '{{title}}'.startswith('"') else '{{title}}'
+subtitle_raw = '''{{subtitle}}'''
+subtitle = json.loads(subtitle_raw) if subtitle_raw.startswith('"') else subtitle_raw if subtitle_raw != '{{' + 'subtitle}}' else ''
+slides_data = json.loads('''{{slides}}''')
+
+prs = Presentation()
+prs.slide_width = Inches(13.333)
+prs.slide_height = Inches(7.5)
+
+# Title slide
+slide = prs.slides.add_slide(prs.slide_layouts[0])
+slide.shapes.title.text = title
+if subtitle and slide.placeholders[1]:
+    slide.placeholders[1].text = subtitle
+
+# Content slides
+for s in slides_data:
+    slide = prs.slides.add_slide(prs.slide_layouts[1])
+    slide.shapes.title.text = s.get('title', '')
+    body = slide.placeholders[1].text_frame
+    body.clear()
+    for i, bullet in enumerate(s.get('bullets', [])):
+        if i == 0:
+            body.paragraphs[0].text = bullet
+        else:
+            p = body.add_paragraph()
+            p.text = bullet
+        body.paragraphs[-1].font.size = Pt(18)
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'presentation.pptx')
+prs.save(filepath)
+print('Generated: presentation.pptx')
+`;
+const SEED_CHART = `import os, json
+import matplotlib
+matplotlib.use('Agg')
+import matplotlib.pyplot as plt
+
+chart_type = '{{chartType}}'
+title = json.loads('''{{title}}''') if '{{title}}'.startswith('"') else '{{title}}'
+labels = json.loads('''{{labels}}''')
+values = json.loads('''{{values}}''')
+xlabel_raw = '''{{xlabel}}'''
+ylabel_raw = '''{{ylabel}}'''
+xlabel = xlabel_raw if xlabel_raw != '{{' + 'xlabel}}' else ''
+ylabel = ylabel_raw if ylabel_raw != '{{' + 'ylabel}}' else ''
+
+fig, ax = plt.subplots(figsize=(10, 6))
+colors = ['#4472C4', '#ED7D31', '#A5A5A5', '#FFC000', '#5B9BD5', '#70AD47', '#264478', '#9B59B6']
+
+if chart_type == 'bar':
+    bars = ax.bar(labels, values, color=colors[:len(labels)])
+    for bar, val in zip(bars, values):
+        ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + max(values)*0.01,
+                f'{val:,.0f}' if isinstance(val, (int, float)) else str(val),
+                ha='center', va='bottom', fontsize=9)
+    if xlabel: ax.set_xlabel(xlabel)
+    if ylabel: ax.set_ylabel(ylabel)
+elif chart_type == 'line':
+    ax.plot(labels, values, marker='o', linewidth=2, color='#4472C4', markersize=6)
+    for i, val in enumerate(values):
+        ax.annotate(f'{val:,.0f}', (labels[i], val), textcoords="offset points",
+                    xytext=(0, 10), ha='center', fontsize=9)
+    if xlabel: ax.set_xlabel(xlabel)
+    if ylabel: ax.set_ylabel(ylabel)
+elif chart_type == 'pie':
+    ax.pie(values, labels=labels, colors=colors[:len(labels)], autopct='%1.1f%%', startangle=90)
+
+ax.set_title(title, fontsize=14, fontweight='bold', pad=15)
+plt.tight_layout()
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'chart.png')
+fig.savefig(filepath, dpi=150, bbox_inches='tight')
+plt.close()
+print('Generated: chart.png')
+`;
+const SEED_INVOICE = `import os, json
+from reportlab.lib.pagesizes import A4
+from reportlab.pdfgen import canvas
+from reportlab.lib.units import cm, mm
+from reportlab.lib import colors
+
+inv_number = json.loads('''{{invoiceNumber}}''') if '{{invoiceNumber}}'.startswith('"') else '{{invoiceNumber}}'
+date = json.loads('''{{date}}''') if '{{date}}'.startswith('"') else '{{date}}'
+company = json.loads('''{{companyName}}''') if '{{companyName}}'.startswith('"') else '{{companyName}}'
+customer = json.loads('''{{customerName}}''') if '{{customerName}}'.startswith('"') else '{{customerName}}'
+address_raw = '''{{customerAddress}}'''
+address = json.loads(address_raw) if address_raw.startswith('"') else address_raw if address_raw != '{{' + 'customerAddress}}' else ''
+items = json.loads('''{{items}}''')
+currency_raw = '''{{currency}}'''
+currency = currency_raw if currency_raw != '{{' + 'currency}}' else '$'
+notes_raw = '''{{notes}}'''
+notes = json.loads(notes_raw) if notes_raw.startswith('"') else notes_raw if notes_raw != '{{' + 'notes}}' else ''
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, f'invoice_{inv_number}.pdf')
+w, h = A4
+c = canvas.Canvas(filepath, pagesize=A4)
+
+# Header
+c.setFont('Helvetica-Bold', 24)
+c.setFillColor(colors.HexColor('#2C3E50'))
+c.drawString(2*cm, h - 2.5*cm, 'INVOICE')
+c.setFont('Helvetica', 10)
+c.setFillColor(colors.HexColor('#7F8C8D'))
+c.drawRightString(w - 2*cm, h - 2.5*cm, f'#{inv_number}')
+c.drawRightString(w - 2*cm, h - 3*cm, f'Date: {date}')
+
+# Company & Customer
+y = h - 4.5*cm
+c.setFillColor(colors.HexColor('#2C3E50'))
+c.setFont('Helvetica-Bold', 11)
+c.drawString(2*cm, y, 'From:')
+c.drawString(10*cm, y, 'To:')
+c.setFont('Helvetica', 10)
+c.setFillColor(colors.black)
+c.drawString(2*cm, y - 0.5*cm, company)
+c.drawString(10*cm, y - 0.5*cm, customer)
+if address:
+    for i, line in enumerate(address.split('\\n')):
+        c.drawString(10*cm, y - (1 + i*0.5)*cm, line)
+
+# Table header
+y = h - 8*cm
+c.setFillColor(colors.HexColor('#4472C4'))
+c.rect(2*cm, y - 0.1*cm, w - 4*cm, 0.7*cm, fill=1, stroke=0)
+c.setFillColor(colors.white)
+c.setFont('Helvetica-Bold', 9)
+c.drawString(2.2*cm, y + 0.1*cm, 'Description')
+c.drawRightString(12*cm, y + 0.1*cm, 'Qty')
+c.drawRightString(15*cm, y + 0.1*cm, 'Unit Price')
+c.drawRightString(w - 2.2*cm, y + 0.1*cm, 'Amount')
+
+# Items
+c.setFillColor(colors.black)
+c.setFont('Helvetica', 9)
+total = 0
+for i, item in enumerate(items):
+    row_y = y - (i + 1) * 0.6*cm
+    qty = item.get('quantity', 0)
+    price = item.get('unitPrice', 0)
+    amount = qty * price
+    total += amount
+    if i % 2 == 1:
+        c.setFillColor(colors.HexColor('#F0F4F8'))
+        c.rect(2*cm, row_y - 0.1*cm, w - 4*cm, 0.6*cm, fill=1, stroke=0)
+        c.setFillColor(colors.black)
+    c.drawString(2.2*cm, row_y + 0.1*cm, str(item.get('description', '')))
+    c.drawRightString(12*cm, row_y + 0.1*cm, str(qty))
+    c.drawRightString(15*cm, row_y + 0.1*cm, f'{currency}{price:,.2f}')
+    c.drawRightString(w - 2.2*cm, row_y + 0.1*cm, f'{currency}{amount:,.2f}')
+
+# Total
+total_y = y - (len(items) + 1.5) * 0.6*cm
+c.setStrokeColor(colors.HexColor('#4472C4'))
+c.line(12*cm, total_y + 0.4*cm, w - 2*cm, total_y + 0.4*cm)
+c.setFont('Helvetica-Bold', 11)
+c.drawRightString(15*cm, total_y, 'Total:')
+c.setFillColor(colors.HexColor('#2C3E50'))
+c.drawRightString(w - 2.2*cm, total_y, f'{currency}{total:,.2f}')
+
+# Notes
+if notes:
+    notes_y = total_y - 2*cm
+    c.setFillColor(colors.HexColor('#7F8C8D'))
+    c.setFont('Helvetica-Bold', 9)
+    c.drawString(2*cm, notes_y, 'Notes:')
+    c.setFont('Helvetica', 9)
+    c.setFillColor(colors.black)
+    for i, line in enumerate(notes.split('\\n')):
+        c.drawString(2*cm, notes_y - (i + 1) * 0.4*cm, line)
+
+c.save()
+print(f'Generated: invoice_{inv_number}.pdf')
+`;
+const SEED_DATA_SUMMARY = `import os, json
+import pandas as pd
+from docx import Document
+from docx.shared import Inches, Pt
+from docx.enum.table import WD_TABLE_ALIGNMENT
+
+title = json.loads('''{{title}}''') if '{{title}}'.startswith('"') else '{{title}}'
+raw_data = json.loads('''{{data}}''')
+group_by_raw = '''{{groupByField}}'''
+group_by = group_by_raw if group_by_raw != '{{' + 'groupByField}}' else None
+sort_by_raw = '''{{sortByField}}'''
+sort_by = sort_by_raw if sort_by_raw != '{{' + 'sortByField}}' else None
+top_n_raw = '''{{topN}}'''
+top_n = int(top_n_raw) if top_n_raw != '{{' + 'topN}}' else 10
+
+df = pd.DataFrame(raw_data)
+doc = Document()
+doc.add_heading(title, 0)
+
+# Overview
+doc.add_heading('Overview', level=1)
+doc.add_paragraph(f'Total records: {len(df)}')
+doc.add_paragraph(f'Fields: {", ".join(df.columns.tolist())}')
+
+# Numeric summary
+num_cols = df.select_dtypes(include='number').columns.tolist()
+if num_cols:
+    doc.add_heading('Numeric Summary', level=1)
+    stats = df[num_cols].describe().round(2)
+    table = doc.add_table(rows=len(stats) + 1, cols=len(num_cols) + 1)
+    table.style = 'Light Grid Accent 1'
+    table.alignment = WD_TABLE_ALIGNMENT.CENTER
+    table.rows[0].cells[0].text = 'Metric'
+    for j, col in enumerate(num_cols):
+        table.rows[0].cells[j + 1].text = col
+    for i, (idx, row) in enumerate(stats.iterrows()):
+        table.rows[i + 1].cells[0].text = str(idx)
+        for j, col in enumerate(num_cols):
+            table.rows[i + 1].cells[j + 1].text = str(row[col])
+
+# Group-by analysis
+if group_by and group_by in df.columns:
+    doc.add_heading(f'Grouped by: {group_by}', level=1)
+    grouped = df.groupby(group_by)
+    summary_rows = []
+    for name, group in grouped:
+        row_info = {'Group': str(name), 'Count': len(group)}
+        for nc in num_cols:
+            row_info[f'{nc} (sum)'] = round(group[nc].sum(), 2)
+            row_info[f'{nc} (avg)'] = round(group[nc].mean(), 2)
+        summary_rows.append(row_info)
+    if summary_rows:
+        headers = list(summary_rows[0].keys())
+        table = doc.add_table(rows=len(summary_rows) + 1, cols=len(headers))
+        table.style = 'Light Grid Accent 1'
+        for j, h in enumerate(headers):
+            table.rows[0].cells[j].text = h
+        for i, sr in enumerate(summary_rows):
+            for j, h in enumerate(headers):
+                table.rows[i + 1].cells[j].text = str(sr[h])
+
+# Top records
+if sort_by and sort_by in df.columns:
+    df_sorted = df.sort_values(sort_by, ascending=False)
+else:
+    df_sorted = df
+top = df_sorted.head(top_n)
+doc.add_heading(f'Top {top_n} Records', level=1)
+cols = top.columns.tolist()
+table = doc.add_table(rows=len(top) + 1, cols=len(cols))
+table.style = 'Light Grid Accent 1'
+for j, col in enumerate(cols):
+    table.rows[0].cells[j].text = str(col)
+for i, (_, row) in enumerate(top.iterrows()):
+    for j, col in enumerate(cols):
+        table.rows[i + 1].cells[j].text = str(row[col])
+
+output_dir = os.environ.get('OUTPUT_DIR', '/output')
+filepath = os.path.join(output_dir, 'summary_report.docx')
+doc.save(filepath)
+print('Generated: summary_report.docx')
+`;
+const SEED_DATA_TRANSFORM = `const fs = require('fs');
+const path = require('path');
+
+const data = {{data}};
+const format = '{{format}}';
+const filename = '{{filename}}' !== '{{' + 'filename}}' ? '{{filename}}' : 'result';
+const outputDir = process.env.OUTPUT_DIR || '/output';
+
+if (format === 'csv') {
+  const headers = Object.keys(data[0] || {});
+  const csv = [
+    headers.join(','),
+    ...data.map(row => headers.map(h => JSON.stringify(row[h] ?? '')).join(','))
+  ].join('\\n');
+  const outPath = path.join(outputDir, filename + '.csv');
+  fs.writeFileSync(outPath, csv, 'utf-8');
+  console.log('Generated: ' + filename + '.csv');
+} else {
+  const outPath = path.join(outputDir, filename + '.json');
+  fs.writeFileSync(outPath, JSON.stringify(data, null, 2), 'utf-8');
+  console.log('Generated: ' + filename + '.json');
+}
+`;
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  SkillManager
+});

package/dist/server/services/SkillRepositoryService.d.ts

@@ -0,0 +1,22 @@
+export interface SkillPackageMetadata {
+    path: string;
+    metadata: Record<string, any>;
+    instructions: string;
+    code: string | null;
+}
+export declare class SkillRepositoryService {
+    private baseDir;
+    constructor(storagePath: string);
+    /**
+     * Extract a zip file to the repository.
+     * Returns parsed metadata from SKILL.md / skill.yaml
+     */
+    extractSkillPackage(skillName: string, zipFilePath: string): Promise<SkillPackageMetadata>;
+    getSkillPath(skillName: string): string;
+    getSkillCode(skillName: string): string | null;
+    copySkillPackageTo(skillName: string, destDir: string): void;
+    readSkillPackage(packageDir: string): SkillPackageMetadata;
+    copyDirectoryTo(srcDir: string, destDir: string): void;
+    private getSkillCodeFromDir;
+    private aggregateOtherMarkdownFiles;
+}