sajalhalder

FuncVul: Detect Vulnerabilities in Code Functions

FuncVul applies a fine-tuned GraphCodeBERT model to code chunks to detect vulnerable functions in C/C++ and Python. It is aimed at researchers and security teams evaluating function-level vulnerability detection, with datasets and baselines included.

View on GitHub Compare
Stars
21
Forks
6
Stars, last 30 days
+0
Last commit
4 months ago
Contributors
1

Our take

Quietno commits in the last 90 days and no release in a year

Health score

16/100

How it is scored
Activity
19
Community
4
Issues
21
Pull requests
20

+0 stars in the last 30 days

  • Fewer contributors than 91% of the projects we track

FuncVul offers a focused research artifact and accompanying datasets, but its quiet maintenance and single-contributor history call for independent validation before relying on it in a security workflow.

Good fit if

  • You are researching code-chunk-based vulnerability detection in C/C++ or Python.
  • You want to examine the included datasets, labeling approaches, and baseline comparisons.
  • You can work with a Python project and assess research code before applying its predictions.

Look elsewhere if

  • You need an actively maintained vulnerability scanning product or a documented deployment service.
  • Your adoption process requires recent releases or evidence of ongoing multi-contributor maintenance.
  • You need a clearly stated license before using or redistributing the code or datasets.
All health signals
Last commit2026-06-05 (4 months ago)
Commits, last 90 days0
Releases, last 12 months0
Contributors1
Issues closed, last 90 days0
Pull requests merged, last 90 days0
Project age16 months

Checked on 2026-10-08 with the GitHub API.

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

20

Forks

6

Change since Nov 14, 2026: 0 stars·0 forks

Overview

FuncVul is a research implementation for detecting vulnerable code at function level in C/C++ and Python. Rather than classify an entire function as one unit, it analyzes smaller code chunks, with the goal of locating vulnerable portions and identifying multiple vulnerabilities within a function.

The repository supports experiments using six datasets built with different chunk formats and vulnerability-labeling approaches. It is most relevant to security researchers and teams studying code vulnerability classifiers, rather than developers seeking a ready-to-deploy scanning service.

Key Features

  • Fine-tunes GraphCodeBERT for function-level vulnerability detection.
  • Uses code chunks, including three-line-based chunks, rather than relying only on full-function input.
  • Targets C/C++ and Python code.
  • Includes six datasets with different chunk types and labeling methods.
  • Provides generic-code chunk variants alongside code chunks.
  • Includes extended chunk-size variants for evaluating how added lines affect results.
  • Provides a script to run FuncVul and three baseline models.

Use Cases

  • Security researchers can reproduce or extend experiments on function-level vulnerability detection.
  • Teams comparing code-chunk and full-function approaches can use the supplied datasets and model implementation as a research baseline.
  • Researchers studying how vulnerability labels affect model results can compare datasets labeled with patch information and LLM assistance against LLM-only labels.
  • Developers investigating vulnerabilities across unfamiliar projects can explore the approach as a model prototype, rather than treating it as a production scanner.

What you need

Detected in the repository

  • Python (from requirements.txt)

Getting Started

With the repository and its Python requirements available, run:

python FuncVul+Baselines.py

See the README for dataset details and research context.

License in Plain Words

No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.

Alternatives

  • opengrep: Opengrep detects security issues with customizable pattern rules across codebases, rather than using a fine-tuned model to classify C/C++ and Python functions.
  • pyre-check: Pyre's Pysa analyzes Python security flows, while FuncVul uses a fine-tuned model to classify vulnerable functions in Python and C/C++.
ProjectLanguageLicenseStarsStatus
FuncVulPython–21Quiet
opengrepOCamlLGPL-2.13.1kActive
pyre-checkOCamlMIT7.2kArchived

Considerations

The repository is marked quiet: it has had no commits in the last 90 days and no release in a year. One contributor accounts for all recorded contributions, and no issues were closed in the last 90 days, so teams should expect limited maintenance signals and review the code for their own needs.

The input identifies Python requirements but does not specify a GPU, database, container setup, CI, or tests. The README also does not provide detailed installation or deployment instructions. No license is listed in the provided repository information, so check the repository for current licensing terms before reuse.

Found this useful?

Share it with someone who would like FuncVul.

Comparisons