happy-notes

happy-notes

An Agent Skills compatible skill for managing iflow knowledge bases through AI agents.

Features

• Knowledge Base Management - Create, list, update, delete, and share notebooks
• File Import - Upload local files (PDF, DOCX, TXT, Markdown, images) and import web URLs
• Content Generation - Generate reports (PDF/DOCX/Markdown), PPTs, podcasts, mind maps, and videos from your knowledge base
• Generation Progress - Check the status of generation tasks
• Web Search & Import - Search the web or academic papers (arXiv, etc.), auto-import results, and generate summaries
• Deep Research - Multi-round web search with comprehensive research report generation
• Semantic Search - Deep content retrieval across your knowledge base
• File Management - List, rename, delete, batch-delete, view details, and retry failed files in notebooks
• Sharing - Generate read-only share links for notebooks

Compatibility

This skill follows the open Agent Skills standard and works with multiple AI agent platforms:

Agent	Install Method
Claude Code	Copy to ~/.claude/skills/happy-notes/ or .claude/skills/happy-notes/
OpenClaw	npx clawhub@latest install happy-notes
Devin	Commit to .agents/skills/happy-notes/ in your repo
Windsurf	Copy to .windsurf/skills/happy-notes/
CrewAI	skills=["./happy-notes"]

For any agent that supports the Agent Skills standard, simply place this skill directory under the agent's skill discovery path.

Setup

1. Get an API Key

Visit iflow.cn to create your API key.

2. Configure Credentials

Choose one of the following methods:

# Option A - Config file (recommended)
mkdir -p ~/.config/happy-notes && echo "your_api_key" > ~/.config/happy-notes/api-key

# Option B - Environment variable
export IFLOW_API_KEY="your_api_key"

The skill resolves credentials in order: environment variable > config file.

Project Structure

happy-notes/
├── SKILL.md                  # Main skill definition (entry point)
├── knowledge-base/
│   └── SKILL.md              # Knowledge base & file management APIs
├── reports/
│   └── SKILL.md              # Content generation APIs
├── search/
│   └── SKILL.md              # Web search APIs
├── scripts/
│   ├── iflow_common.py       # Shared utilities (auth, API helpers, retry logic)
│   ├── pipeline_create_kb_and_generate.py    # Pipeline 1: Create KB + upload + generate
│   ├── pipeline_search_and_generate.py       # Pipeline 2: Search KB + generate
│   ├── pipeline_import_and_generate.py       # Pipeline 3: Import to existing KB + generate
│   ├── pipeline_semantic_search.py           # Pipeline 4: Semantic search + generate/share
│   ├── pipeline_file_management.py           # Pipeline 5: File list/rename/delete/info/retry
│   ├── pipeline_web_search.py                # Pipeline 6: Web search + import + generate
│   ├── pipeline_generate.py                  # Direct generation on existing KB
│   ├── pipeline_kb.py                        # Knowledge base CRUD management
│   ├── pipeline_check_status.py              # Check generation task progress
│   └── pipeline_share.py                     # Share knowledge base
├── examples/                 # Usage examples for common workflows
│   ├── student-literature-review.md
│   ├── add-file-then-generate.md
│   ├── search-then-generate.md
│   ├── semantic-search.md
│   ├── text-import.md
│   ├── ppt-with-preset.md
│   ├── share-knowledge-base.md
│   ├── long-task-async.md
│   ├── web-search-fast.md
│   └── web-search-deep.md
└── references/
    ├── api.md                # Full API reference
    ├── pipelines.md          # Pipeline parameter details & output formats
    └── kb-matching.md        # Knowledge base matching logic

Quick Examples

Create a knowledge base, upload files, and generate a report

You: Create a KB called "Research Papers", upload these PDFs, and generate a literature review.

The agent runs Pipeline 1 behind the scenes:

python3 scripts/pipeline_create_kb_and_generate.py \
  --name "Research Papers" \
  --files "/path/to/paper1.pdf,/path/to/paper2.pdf" \
  --output-type "PDF" \
  --query "Generate a literature review comparing methodologies"

Search the web for papers and import into your KB

You: Search for recent papers on LLM agents and create a summary report.

The agent runs Pipeline 6:

python3 scripts/pipeline_web_search.py \
  --kb "Research Papers" \
  --query "LLM agent latest research" \
  --source SCHOLAR \
  --output-type PDF

Save a quick note

You: Record that I spent $50 on lunch today.

The agent auto-matches (or creates) the right notebook and imports via Pipeline 3.

Share a knowledge base

You: Share the "Research Papers" notebook with my team.

The agent generates a read-only share link.

Supported Output Types

Type	Description
PDF	PDF report
DOCX	Word document
MARKDOWN	Markdown document
PPT	PowerPoint presentation (supports style presets)
XMIND	Mind map
PODCAST	Audio podcast
VIDEO	Video content

License

MIT