# Docx Read Fallback

> Use run_shell with python-docx as reliable fallback when read_file fails on .docx files

- **Type:** Skill
- **Install:** `agentstack add skill-hkuds-openspace-docx-read-fallback`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [HKUDS](https://agentstack.voostack.com/s/hkuds)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [HKUDS](https://github.com/HKUDS)
- **Source:** https://github.com/HKUDS/OpenSpace/tree/main/gdpval_bench/skills/docx-read-fallback

## Install

```sh
agentstack add skill-hkuds-openspace-docx-read-fallback
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# DOCX Read Fallback

When `read_file` or `execute_code_sandbox` fails to read `.docx` files, use `run_shell` with python-docx as a reliable workaround.

## When to Use

- `read_file` fails, times out, or returns errors on `.docx` files
- `execute_code_sandbox` attempts to read the docx fail
- You need to extract text content from a Word document
- Multiple standard approaches have been exhausted

## How to Use

### Basic Text Extraction

```bash
python -c "import docx; doc = docx.Document('path/to/file.docx'); print('\n'.join([p.text for p in doc.paragraphs]))"
```

### Using run_shell Tool

```
run_shell command="python -c \"import docx; doc = docx.Document('path/to/file.docx'); print('\n'.join([p.text for p in doc.paragraphs]))\"" timeout=60
```

### Extract Paragraphs with Indices

```bash
python -c "import docx; doc = docx.Document('file.docx'); [print(f'P{i}: {p.text}') for i, p in enumerate(doc.paragraphs) if p.text.strip()]"
```

### Extract Tables

```bash
python -c "import docx; doc = docx.Document('file.docx'); [[print([[cell.text for cell in row.cells] for row in table.rows]) for table in doc.tables]]"
```

### Extract Headings (by style)

```bash
python -c "import docx; doc = docx.Document('file.docx'); [print(p.text) for p in doc.paragraphs if p.style.name.startswith('Heading')]"
```

## Prerequisites

Ensure python-docx is available:

```bash
python -c "import docx; print('docx available')"
```

If not installed:

```bash
pip install python-docx
```

## Tips

- Use absolute paths to avoid working directory issues
- Set appropriate `timeout` (30-60 seconds for large documents)
- Escape quotes properly when embedding in shell commands
- For large documents, extract content in chunks or filter by paragraph index
- This approach bypasses file type detection issues in read_file

## Example Workflow

1. Try `read_file` on the .docx file
2. If it fails, verify python-docx availability
3. Use `run_shell` with the python-docx extraction command
4. Parse the stdout to get document content
5. Proceed with your analysis using the extracted text

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [HKUDS](https://github.com/HKUDS)
- **Source:** [HKUDS/OpenSpace](https://github.com/HKUDS/OpenSpace)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-hkuds-openspace-docx-read-fallback
- Seller: https://agentstack.voostack.com/s/hkuds
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.