|
| 1 | +--- |
| 2 | +title: Python File Audit |
| 3 | +short_title: Python File Audit |
| 4 | +--- |
| 5 | + |
| 6 | +Handling untrusted input files is one of the most common vectors for security vulnerabilities in Python applications. Accepting file uploads or processing third-party data without explicit validation violates zero-trust principles and exposes your system to Denial of Service (DoS), Remote Code Execution (RCE), and Arbitrary File Write attacks. |
| 7 | + |
| 8 | +The [`fileaudit`](https://github.com/nocomplexity/fileaudit) library provides a lightweight, [open-source](https://nocomplexity.github.io/fileaudit/license.html#software-license) defense layer designed to intercept and validate files before your application processes them. |
| 9 | + |
| 10 | +## File-Based Threat Vectors |
| 11 | + |
| 12 | +When your application handles untrusted input, the attack surface varies depending on the archive or serialisation format. |
| 13 | + |
| 14 | +### Archive Exploits (ZIP, TAR, GZ) |
| 15 | + |
| 16 | +- **Archive / Decompression Bombs:** A heavily compressed file expands exponentially upon extraction (e.g., a few kilobytes expanding into hundreds of gigabytes), quickly exhausting disk space and memory. |
| 17 | +- **Inode & DoS Bombs:** Archives containing millions of microscopic files designed to consume all available filesystem inodes or exhaust CPU during unpacking. |
| 18 | +- **Path Traversal (Zip Slip):** Archive entries containing relative paths (e.g., `../../etc/passwd`) designed to overwrite system or application files outside the extraction directory. |
| 19 | +- **Symlink & Hardlink Attacks:** Archives embedded with symbolic or hard links targeting sensitive host paths, allowing subsequent file writes to redirect to critical system locations. |
| 20 | +- **Device Node Injection:** Archives containing character devices, block devices, or FIFOs (named pipes) intended to cause hangs or facilitate privilege escalation. |
| 21 | +- **Directory Traversal Depths:** Deeply nested folder structures intended to trigger recursive extraction stack overflows or degrade filesystem operations. |
| 22 | + |
| 23 | +### XML Processing Exploits |
| 24 | + |
| 25 | +Processing untrusted XML documents using standard parsers exposes the application to severe parser-level attacks: |
| 26 | + |
| 27 | +- **Billion Laughs (Entity Expansion):** Exponential inline DTD entity expansion that consumes all available RAM in seconds. |
| 28 | +- **XML External Entity (XXE):** Misconfigured DTD parsers disclosing local system files or making SSRF calls via URI handlers. |
| 29 | +- **Attribute & Structure Bombs:** Elements containing thousands of attributes or extreme nesting depths designed to trigger stack overflows and CPU starvation. |
| 30 | + |
| 31 | +### JSON & Structured Data Exploits |
| 32 | + |
| 33 | +As highlighted in the [official Python documentation](https://docs.python.org/3/library/json.html), native decoders can be vulnerable to resource exhaustion when parsing malformed or excessively large payloads: |
| 34 | + |
| 35 | +:::{warning} |
| 36 | +Unbounded JSON strings or deeply nested objects can force `json.loads()` to consume excessive CPU cycles and RAM. Enforcing strict size and recursion limits prior to parsing is required when dealing with untrusted origins. |
| 37 | +::: |
| 38 | + |
| 39 | +## Defending Files with `fileaudit` |
| 40 | + |
| 41 | +The `fileaudit` library mitigates these risks by applying safe extraction constraints, path sanitisation, and size boundaries prior to processing. |
| 42 | + |
| 43 | +### Supported Formats |
| 44 | + |
| 45 | +| Extension | Target File Format | Primary Enforced Protections | |
| 46 | +| :--- | :--- | :--- | |
| 47 | +| `.csv` | Tabular Data | Length limits, delimiter boundaries | |
| 48 | +| `.gz`, `.tgz`, `.tar.gz` | Compressed Archives | Decompression ratio caps, size limits | |
| 49 | +| `.json` | JSON Documents | String length, structure depth, raw size | |
| 50 | +| `.py` | Python Source Files | File size, static path checks | |
| 51 | +| `.tar` | Tape Archives | Member counts, symlink/FIFO rejection | |
| 52 | +| `.xml` | XML Documents | Entity expansion bounds, node depth limits | |
| 53 | +| `.zip` | ZIP Archives | Path traversal blocking, size limits | |
| 54 | + |
| 55 | +### Key Protections Enforced |
| 56 | + |
| 57 | +:::{admonition} Built-in Security Controls |
| 58 | +:class: tip |
| 59 | + |
| 60 | +- **Decompression Caps:** Monitors GZip expansion ratios to catch decompression bombs early. |
| 61 | +- **Archive Size & Member Bounds:** Imposes hard upper limits on total extracted size, entry counts, and individual file sizes. |
| 62 | +- **Path & Type Sanitisation:** Automatically strips path traversal sequences (`../`), rejecting symlinks, hard links, FIFOs, and special device nodes. |
| 63 | +- **Structural Bounds:** Restricts maximum filename lengths, total directory depth, and nesting levels. |
| 64 | +::: |
| 65 | + |
| 66 | +## Implementation |
| 67 | + |
| 68 | +### Installation |
| 69 | + |
| 70 | +Install [`fileaudit`](https://github.com/nocomplexity/fileaudit) from [PyPI](https://pypi.org/project/fileaudit/): |
| 71 | + |
| 72 | +```bash |
| 73 | +pip install fileaudit |
| 74 | + |
| 75 | +``` |
| 76 | + |
| 77 | +### Usage Modes |
| 78 | + |
| 79 | +Each format validator (such as `validate_tar_gz`, `validate_zip`, or `validate_json`) supports two integration patterns depending on your architecture. |
| 80 | + |
| 81 | +#### 1. Decorator Mode |
| 82 | + |
| 83 | +Wrap processing functions directly. The validator inspects the input parameter and aborts execution before the function body runs if checks fail: |
| 84 | + |
| 85 | +```python |
| 86 | +from fileaudit import validate_tar_gz |
| 87 | + |
| 88 | +@validate_tar_gz |
| 89 | +def process_upload(file_path: str) -> None: |
| 90 | + # Function executes only if the archive passes all security bounds |
| 91 | + ... |
| 92 | + |
| 93 | +``` |
| 94 | + |
| 95 | +#### 2. Direct Functional Call |
| 96 | + |
| 97 | +Perform explicit programmatic verification within your validation pipeline: |
| 98 | + |
| 99 | +```python |
| 100 | +from fileaudit import validate_tar_gz |
| 101 | + |
| 102 | +def handle_incoming_file(file_path: str) -> None: |
| 103 | + if not validate_tar_gz(file_path): |
| 104 | + raise ValueError("File failed security audit checks.") |
| 105 | + |
| 106 | + # Safe to proceed with extraction |
| 107 | + ... |
| 108 | + |
| 109 | +``` |
| 110 | + |
| 111 | +For advanced configuration, parameter tuning, and custom threshold adjustments, consult the official [Python File Audit Documentation](https://nocomplexity.github.io/fileaudit/intro.html). |
| 112 | + |
0 commit comments