Implementation:Trailofbits Fickling Check For Corruption

Knowledge Sources	Fickling
Domains	Security, File_Format, Integrity
Last Updated	2026-02-14 14:00 GMT

Overview

Concrete tool for detecting structural corruption in PyTorch files provided by the Fickling library.

Description

check_for_corruption takes the properties dictionary from find_file_properties and applies structural integrity rules. Currently checks for TorchScript files with model.json but missing both attributes.pkl and constants.pkl. Returns a tuple of (corrupted: bool, reason: str).

Usage

Use after find_file_properties to verify file structural integrity before format identification or analysis.

Code Reference

Source Location

• Repository: fickling
• File: fickling/polyglot.py
• Lines: L259-273

Signature

def check_for_corruption(properties: dict) -> tuple[bool, str]:
    """Check for structural corruption in PyTorch files.

    Args:
        properties: Properties dict from find_file_properties().

    Returns:
        Tuple of (corrupted: bool, reason: str).
        Returns (False, "") if no corruption detected.
    """

Import

from fickling.polyglot import check_for_corruption

I/O Contract

Inputs

Name	Type	Required	Description
properties	dict	Yes	Properties dictionary from find_file_properties()

Outputs

Name	Type	Description
(return)	tuple[bool, str]	(corrupted, reason); (False, "") if no corruption

Usage Examples

Check File Integrity

from fickling.polyglot import find_file_properties, check_for_corruption

properties = find_file_properties("model.pt")
corrupted, reason = check_for_corruption(properties)

if corrupted:
    print(f"CORRUPTED: {reason}")
else:
    print("File structure is valid")

Related Pages

Implements Principle

• Principle:Trailofbits_Fickling_File_Corruption_Detection