Integrated Report Bundle logo

Integrated Report Bundle

OrganizationPopular
HKUDS
integrated-report-bundle

Generate spreadsheet, diagram, and PDF report as a cohesive deliverable set with iteration budgeting per component

Overview

PublisherHKUDS
RepositoryOpenSpace
Skill nameintegrated-report-bundle
Stars
7.8K
Forks
923
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by HKUDS on GitHub. Read the source before you install it.

Installation

Install the Integrated Report Bundle AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/HKUDS/OpenSpace/tree/main/benchmarks/gdpval/skills/pdf-checklist-workflow-enhanced-a53b26
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/HKUDS/OpenSpace.git /tmp/OpenSpace
mkdir -p .claude/skills
cp -r /tmp/OpenSpace/benchmarks/gdpval/skills/pdf-checklist-workflow-enhanced-a53b26 .claude/skills/integrated-report-bundle
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Integrated Report Bundle in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Integrated Report Bundle on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Integrated Report Bundle is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Integrated Report Bundle Workflow

This skill provides a reusable pattern for creating a complete report package consisting of three interconnected deliverables: a spreadsheet (data table), a diagram (visual layout), and a PDF report (summary document). Designed for resilience with explicit iteration budgeting and workspace management.

Overview

Use this workflow when you need to:

  • Generate multiple interconnected deliverables in a single task
  • Create data spreadsheets with structured tables
  • Produce visual diagrams (PNG/SVG) for layouts or workflows
  • Generate summary PDF reports with scoring or assessments
  • Manage iteration budgets across multiple components
  • Handle sandbox execution failures gracefully

Critical: Iteration Budget Allocation

Total budget must be divided across deliverables:

DeliverableRecommended BudgetPurpose
Spreadsheet8-10 iterationsData gathering, table creation, validation
Diagram8-10 iterationsVisual generation, layout refinement
PDF Report8-10 iterationsReport assembly, formatting, final output
Buffer2-4 iterationsError recovery, verification, fixes

Never spend entire budget on one deliverable. Track iterations explicitly.

Step-by-Step Instructions

Phase 0: Workspace Preparation (1 iteration)

Before generating any deliverables, establish clean workspace state:

bash
# Check current workspace state
ls -la /workspace/

# Create dedicated output directory
mkdir -p /workspace/deliverables

# Verify write permissions
touch /workspace/deliverables/test.tmp && rm /workspace/deliverables/test.tmp

Why: Prevents workspace path confusion that caused previous failures.

Phase 1: Spreadsheet Creation (8-10 iterations max)

Step 1.1: Choose Approach
ModeToolUse When
Aexecute_code_sandbox with openpyxl/pandasProgrammatic table generation
Bshell_agent with Python scriptComplex data processing
Ccreate_file (CSV/XLSX)Simple tabular data
Step 1.2: Generate Spreadsheet Code

Example using openpyxl:

python
from openpyxl import Workbook
from openpyxl.styles import Font, PatternFill, Alignment

wb = Workbook()
ws = wb.active
ws.title = "Hardware Selection"

# Headers with styling
headers = ["Component", "Model", "Qty", "Unit Price", "Total"]
for col, header in enumerate(headers, 1):
    cell = ws.cell(row=1, column=col, value=header)
    cell.font = Font(bold=True)
    cell.fill = PatternFill(start_color="4472C4", fill_type="solid")
    cell.font = Font(color="FFFFFF", bold=True)

# Data rows
data = [
    ["CNC Controller", "Fanuc 0i-MF", 1, 2500, 2500],
    ["Servo Motor", "SGM7G-09A", 4, 450, 1800],
    ["HMI Panel", "GT2110-WTBD", 1, 380, 380]
]

for row_idx, row_data in enumerate(data, 2):
    for col_idx, value in enumerate(row_data, 1):
        ws.cell(row=row_idx, column=col_idx, value=value)

# Total row
ws.cell(row=len(data)+2, column=4, value="TOTAL:")
ws.cell(row=len(data)+2, column=5, value=f"=E2:E{len(data)+1}")

# Save
output_path = '/workspace/deliverables/hardware_selection_table.xlsx'
wb.save(output_path)
print(f'ARTIFACT_PATH:{output_path}')
print('Spreadsheet created successfully')
Step 1.3: Execute and Verify
execute_code_sandbox
  code: <spreadsheet generation code>
  language: python

Verification (use 1 iteration):

bash
run_shell
  command: ls -lh /workspace/deliverables/*.xlsx && file /workspace/deliverables/*.xlsx

If execute_code_sandbox fails with '[ERROR] unknown error':

  1. Retry once with shell_agent instead
  2. If still failing, fall back to create_file with CSV format
  3. Document the failure and proceed to next deliverable

Iteration checkpoint: After 8 iterations on spreadsheet, move to diagram regardless of completion status.

Phase 2: Diagram Generation (8-10 iterations max)

Step 2.1: Choose Diagram Type
TypeToolBest For
Flowchartgraphviz + pydotProcess flows, decision trees
Layoutmatplotlib + patchesPhysical layouts, workcells
Architecturediagrams librarySystem architecture
Simple shapesPIL/PillowBasic diagrams
Step 2.2: Generate Diagram Code

Example using matplotlib for layout diagram:

python
import matplotlib.pyplot as plt
import matplotlib.patches as patches

fig, ax = plt.subplots(figsize=(12, 8))
ax.set_xlim(0, 100)
ax.set_ylim(0, 80)
ax.set_aspect('equal')
ax.axis('off')

# Title
ax.text(50, 75, 'CNC Workcell Layout', fontsize=16, ha='center', 
        fontweight='bold', bbox=dict(boxstyle='round', facecolor='wheat'))

# CNC Machine
cnc = patches.Rectangle((10, 30), 25, 20, linewidth=2, 
                        edgecolor='darkblue', facecolor='lightblue',
                        label='CNC Machine')
ax.add_patch(cnc)
ax.text(22.5, 40, 'CNC\nMachine', ha='center', va='center', fontsize=10)

# Robot Arm
robot = patches.Circle((50, 40), 8, linewidth=2,
                       edgecolor='darkgreen', facecolor='lightgreen',
                       label='Robot')
ax.add_patch(robot)
ax.text(50, 40, 'Robot', ha='center', va='center', fontsize=10)

# Material Storage
storage = patches.Rectangle((70, 30), 20, 20, linewidth=2,
                           edgecolor='darkred', facecolor='lightcoral',
                           label='Storage')
ax.add_patch(storage)
ax.text(80, 40, 'Material\nStorage', ha='center', va='center', fontsize=9)

# Arrows showing flow
ax.annotate('', xy=(35, 40), xytext=(42, 40),
           arrowprops=dict(arrowstyle='->', color='gray', lw=2))
ax.annotate('', xy=(58, 40), xytext=(65, 40),
           arrowprops=dict(arrowstyle='->', color='gray', lw=2))

# Legend
ax.legend(loc='upper right', bbox_to_anchor=(0.98, 0.98))

# Save
output_path = '/workspace/deliverables/cnc_workcell_layout.png'
plt.savefig(output_path, dpi=150, bbox_inches='tight')
print(f'ARTIFACT_PATH:{output_path}')
print('Diagram created successfully')
Step 2.3: Execute and Verify
execute_code_sandbox
  code: <diagram generation code>
  language: python

Verification:

bash
run_shell
  command: ls -lh /workspace/deliverables/*.png && file /workspace/deliverables/*.png

If execute_code_sandbox fails:

  1. First retry: Simplify the diagram (remove complex elements)
  2. Second retry: Use shell_agent with explicit matplotlib installation
  3. Third retry: Create ASCII/text-based diagram as fallback
  4. Document and proceed (diagram is often optional compared to PDF)

Iteration checkpoint: After 8 iterations on diagram, move to PDF regardless of completion status.

Phase 3: PDF Report Generation (8-10 iterations max)

Step 3.1: Choose PDF Library
LibraryBest ForInstallation
fpdf2Simple reports, quick setup!pip install fpdf2
reportlabProfessional reports, tables!pip install reportlab
matplotlibChart-heavy reportsBuilt-in
Step 3.2: Generate PDF Code

Example using fpdf2 (simplest, most reliable):

python
from fpdf import FPDF

class ReportPDF(FPDF):
    def header(self):
        self.set_font('Arial', 'B', 14)
        self.cell(0, 10, 'Project Assessment Report', 0, 1, 'C')
        self.line(10, 20, 200, 20)
        self.ln(10)
    
    def footer(self):
        self.set_y(-15)
        self.set_font('Arial', 'I', 8)
        self.cell(0, 10, f'Page {self.page_no()}', 0, 0, 'C')
    
    def section(self, title):
        self.set_font('Arial', 'B', 12)
        self.set_fill_color(200, 220, 255)
        self.cell(0, 8, title, 0, 1, 'L', 1)
        self.ln(3)
    
    def add_table_row(self, col1, col2, col3):
        self.set_font('Arial', '', 10)
        self.cell(60, 8, col1, 1)
        self.cell(80, 8, col2, 1)
        self.cell(40, 8, col3, 1)
        self.ln()

pdf = ReportPDF()
pdf.add_page()
pdf.set_auto_page_break(auto=True, margin=15)

# Executive Summary
pdf.section('Executive Summary')
pdf.set_font('Arial', '', 11)
pdf.multi_cell(0, 7, 'This report provides a comprehensive assessment of the CNC workcell implementation project, including hardware selection, layout design, and cost analysis.')
pdf.ln(5)

# Hardware Selection Summary
pdf.section('Hardware Selection')
pdf.add_table_row('Component', 'Model', 'Cost')
pdf.add_table_row('CNC Controller', 'Fanuc 0i-MF', '$2,500')
pdf.add_table_row('Servo Motors', 'SGM7G-09A (x4)', '$1,800')
pdf.add_table_row('HMI Panel', 'GT2110-WTBD', '$380')
pdf.ln(5)

# Assessment Scores
pdf.section('Assessment Scores')
pdf.add_table_row('Criteria', 'Description', 'Score')
pdf.add_table_row('Completeness', 'All requirements addressed', '8/10')
pdf.add_table_row('Feasibility', 'Technical implementation possible', '9/10')
pdf.add_table_row('Cost Efficiency', 'Within budget constraints', '7/10')
pdf.ln(5)

# Conclusion
pdf.section('Conclusion')
pdf.multi_cell(0, 7, 'The proposed CNC workcell meets all critical requirements with a recommended implementation approach. Total estimated cost: $4,680.')

# Save
output_path = '/workspace/deliverables/assessment_report.pdf'
pdf.output(output_path)
print(f'ARTIFACT_PATH:{output_path}')
print('PDF report created successfully')
Step 3.3: Execute and Verify
execute_code_sandbox
  code: <PDF generation code>
  language: python

Verification (critical - use read_file):

read_file
  filetype: pdf
  file_path: /workspace/deliverables/assessment_report.pdf

Alternative verification:

bash
run_shell
  command: ls -lh /workspace/deliverables/*.pdf && pdftk /workspace/deliverables/*.pdf dump_data 2>/dev/null | head -20

Phase 4: Final Bundle Verification (2-4 iterations)

Verify all deliverables exist and are valid:

bash
run_shell
  command: |
    echo "=== Deliverable Bundle Verification ==="
    echo ""
    echo "Spreadsheet:"
    ls -lh /workspace/deliverables/*.xlsx 2>/dev/null || echo "  NOT FOUND"
    echo ""
    echo "Diagram:"
    ls -lh /workspace/deliverables/*.png 2>/dev/null || echo "  NOT FOUND"
    echo ""
    echo "PDF Report:"
    ls -lh /workspace/deliverables/*.pdf 2>/dev/null || echo "  NOT FOUND"
    echo ""
    echo "=== Bundle Complete ==="

Checklist:

  • At least 2 of 3 deliverables created
  • All files > 1KB (not empty)
  • ARTIFACT_PATH outputs visible in execution logs
  • No unresolved execute_code_sandbox errors

Error Handling Matrix

Error TypeDetectionRecovery ActionMax Retries
[ERROR] unknown error (execute_code_sandbox)Error in tool outputSwitch to shell_agent or simplify code2
File not foundVerification step failsCheck path, re-run generation1
Library not installedImportError in outputAdd !pip install at code start1
Empty file (0 bytes)ls -lh shows 0 sizeRegenerate with error handling1
Iteration budget exceededIteration count reachedDocument partial completion, proceedN/A

Workspace Path Convention

Always use: /workspace/deliverables/filename.extension

Never use:

  • /workspace/filename.extension (root clutter)
  • ./filename.extension (ambiguous)
  • filename.extension (missing path)

Template: Quick Start All Three Deliverables

Use this template structure to generate all three deliverables efficiently:

python
# DELIVERABLE 1: Spreadsheet
import openpyxl
wb = openpyxl.Workbook()
ws = wb.active
ws['A1'] = 'Header'
wb.save('/workspace/deliverables/data.xlsx')
print('ARTIFACT_PATH:/workspace/deliverables/data.xlsx')

# DELIVERABLE 2: Diagram  
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
ax.text(0.5, 0.5, 'Diagram')
plt.savefig('/workspace/deliverables/diagram.png')
print('ARTIFACT_PATH:/workspace/deliverables/diagram.png')

# DELIVERABLE 3: PDF
from fpdf import FPDF
pdf = FPDF()
pdf.add_page()
pdf.set_font('Arial', '', 12)
pdf.cell(0, 10, 'Report')
pdf.output('/workspace/deliverables/report.pdf')
print('ARTIFACT_PATH:/workspace/deliverables/report.pdf')

Iteration Tracking Template

Track your iteration usage explicitly:

Iteration Log:
- Iteration 1-2: Workspace preparation
- Iteration 3-10: Spreadsheet creation (8 used)
- Iteration 11-18: Diagram generation (8 used)
- Iteration 19-26: PDF report (8 used)
- Iteration 27-30: Bundle verification (4 buffer)
Total: 30 iterations

Best Practices

  1. Budget discipline - Never exceed 10 iterations on any single deliverable
  2. Early verification - Verify each deliverable before moving to next phase
  3. Graceful degradation - If one deliverable fails, complete the others
  4. Consistent paths - Always use /workspace/deliverables/ prefix
  5. ARTIFACT_PATH output - Essential for download capability
  6. Error logging - Document failures and recovery actions taken
  7. Fallback strategies - Have alternative approaches ready for each deliverable type

Related Tools

  • execute_code_sandbox - Primary tool for all code generation
  • shell_agent - Fallback when execute_code_sandbox fails repeatedly
  • read_file - Verify PDF content (filetype: pdf, xlsx, png)
  • run_shell - Check file existence, sizes, and workspace state
  • create_file - Fallback for simple data files
  • list_dir - Verify deliverables directory contents

Anti-Patterns to Avoid

  1. Don't spend all iterations on one deliverable
  2. Don't ignore execute_code_sandbox errors - address immediately
  3. Don't use inconsistent paths across deliverables
  4. Don't skip verification steps
  5. Don't assume a deliverable succeeded without checking

Success Criteria

A successful execution produces:

  • Minimum 2 of 3 deliverables (spreadsheet, diagram, PDF)
  • All deliverables in /workspace/deliverables/
  • Valid, non-empty files (>1KB each)
  • Clear ARTIFACT_PATH outputs in execution logs
  • Iteration budget not exceeded

*** End Files

Frequently asked questions

What does the Integrated Report Bundle AI skill do?

Generate spreadsheet, diagram, and PDF report as a cohesive deliverable set with iteration budgeting per component

Why use Integrated Report Bundle on TypingMind?

Because you install it once and use it with any model. Integrated Report Bundle is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Integrated Report Bundle in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/HKUDS/OpenSpace/tree/main/benchmarks/gdpval/skills/pdf-checklist-workflow-enhanced-a53b26. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Integrated Report Bundle?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Integrated Report Bundle?

As many as you like. As long as a model supports skills, you can use Integrated Report Bundle with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Integrated Report Bundle AI skill free?

Yes. It is published on GitHub by HKUDS under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇