* Package name : sys-scan-graph Version : 6.0.0 Upstream Author : Joseph Mazzini jmazzini541@gmail.com * URL : https://github.com/J-mazz/sys-scan-graph * License : Apache-2.0 Programming Lang: C++, Python Description : System security scanner and graph generator sys-scan-graph is a comprehensive system security scanning tool that combines a high-performance C++20 scanning engine with an AI-powered intelligence layer to deliver in-depth Linux system security analysis. The tool performs vulnerability assessments, compliance checks, and provides detailed security insights for Linux systems. The C++ core scanner performs fast, deterministic security enumeration across 16 specialized domains, including process analysis, network socket detection, kernel hardening checks, and SUID/SGID binary tracking. It emphasizes determinism by producing RFC 8785 canonical JSON output, ensuring reproducible results. The Python-based intelligence layer transforms the raw scan data into actionable security intelligence. It utilizes an embedded, fine-tuned Mistral-7B model for analysis, which allows for offline operation and ensures data sovereignty. This layer provides features such as MITRE ATT&CK correlation, baseline anomaly detection, and compliance mapping for standards like PCI DSS 4.0, HIPAA, and NIST CSF 2.0. The Python component is also available on PyPI as `sys-scan-agent`. The project is actively maintained, has comprehensive test coverage with over 900 test cases, and is well-documented. I will be maintaining this package.
Dear Bastian, Maintainers,
Thank you again for the feedback on the RFS for sys-scan-graph (Bug
#1118183).
I have addressed the points previously raised:
1. **Separate Packaging Repository:** The Debian packaging files have been
moved out of the upstream source repository and are now maintained
separately here:
* `https://github.com/J-mazz/sys-scan-debian_package`
<https://github.com/J-mazz/sys-scan-debian_package>
2. **Changelog:** The `debian/changelog` file has been corrected to
contain only the entry for the initial release (`6.0.0-1`), and it now
references the ITP bug #1118183.
For further context on the package:
* **Upstream Source Repository:** The main C++ scanner and Python agent
source code resides at:
* `https://github.com/J-mazz/sys-scan-graph`
<https://github.com/J-mazz/sys-scan-graph>
* **PyPI Package:** The Python intelligence layer component is also
available on PyPI under the name `sys-scan-agent`. The Debian package will
ensure correct integration and dependency handling within the Debian
ecosystem.
* **User Interface:** The primary user interface is command-line based. The
Python intelligence layer can generate comprehensive, self-contained HTML
reports for visualizing scan results and analysis. There is also a separate
GTKmm-based graphical UI (`sys-scan-ui`) in development:
* `https://github.com/J-mazz/sys-scan-ui`
<https://github.com/J-mazz/sys-scan-ui>
* This UI interacts with the `sys-scan-agent` via IPC (multiprocessing
queues) and allows users to view findings and provide feedback.
* **Packaging Plan:** To keep the core `sys-scan-graph` package focused
and lightweight, the plan is to package `sys-scan-ui` separately *after*
the core package has been accepted into Debian. This current RFS pertains
only to the core `sys-scan-graph` (CLI scanner and Python agent).
I believe this addresses the request for more information. Please let me
know if any further details or changes are required.
Regards,
Joseph Mazzini
<jmazzini541@gmail.com>
Hello mentors, I have migrated sys-scan-graph debian package source to: https://salsa.debian.org/J-mazz/sys-scan-debian_package Let me know if there is anything further to be done. regards, Joseph Mazzini Jmazzini541@gmail.com joseph@mazzlabs.works
Hello, This email provides an update for the Intent to Package report, Bug#1118183, concerning sys-scan-graph. I have successfully migrated the project's source repository to Debian's Salsa platform. The new canonical Git repository URL is now: https://salsa.debian.org/J-mazz/sys-scan-graph.git Please use this location for all future reference regarding the packaging of this software. Thank you, Joseph Mazzini <jmazzini541@gmail.com> <joseph@mazzlabs.works>
Dear Mentors, I am writing to formally request that my contact email address for this Intent to Package be updated to joseph@mazzlabs.works; please utilize this address for all future correspondence regarding this submission and the associated package maintenance. I have pushed the latest upstream source code revisions to the primary repository located at https://salsa.debian.org/J-mazz/sys-scan-graph. This update notably integrates both a fine-tuned ML feature and a Qt-based Graphical User Interface; this interface is designed to facilitate Human-in-the-Loop (HitL) functionality, allowing for interactive user oversight during the scanning process. The specific Debian packaging files continue to be maintained in their dedicated repository at https://salsa.debian.org/J-mazz/sys-scan-debian_package. In support of the package's machine learning capabilities, I have also uploaded the model fine-tuning workflows and synthetic dataset curation scripts to https://salsa.debian.org/J-mazz/sys-scan-agent_MLops.git. The specific dataset utilized for this implementation has been published to the Hugging Face hub at https://huggingface.co/datasets/jmazz/sys-scan_synthetic_dataset_v2 to ensure transparency and reproducibility of the model's behavior. Thank you for your attention to these updates. Regards, Joseph Mazzini joseph@mazzlabs.works
Dear Mentors, I am writing to formally request that my contact email address for this Intent to Package be updated to joseph@mazzlabs.works; please utilize this address for all future correspondence regarding this submission and the associated package maintenance. I have pushed the latest upstream source code revisions to the primary repository located at https://salsa.debian.org/J-mazz/sys-scan-graph. This update notably integrates both a fine-tuned ML feature and a Qt-based Graphical User Interface; this interface is designed to facilitate Human-in-the-Loop (HitL) functionality, allowing for interactive user oversight during the scanning process. The specific Debian packaging files continue to be maintained in their dedicated repository at https://salsa.debian.org/J-mazz/sys-scan-debian_package. In support of the package's machine learning capabilities, I have also uploaded the model fine-tuning workflows and synthetic dataset curation scripts to https://salsa.debian.org/J-mazz/sys-scan-agent_MLops.git. The specific dataset utilized for this implementation has been published to the Hugging Face hub at https://huggingface.co/datasets/jmazz/sys-scan_synthetic_dataset_v2 to ensure transparency and reproducibility of the model's behavior. Thank you for your attention to these updates. Regards, Joseph Mazzini joseph@mazzlabs.works
Dear Mentors, I am writing to formally request that my contact email address for this Intent to Package be updated to joseph@mazzlabs.works; please utilize this address for all future correspondence regarding this submission and the associated package maintenance. I have pushed the latest upstream source code revisions to the primary repository located at https://salsa.debian.org/J-mazz/sys-scan-graph. This update notably integrates both a fine-tuned ML feature and a Qt-based Graphical User Interface; this interface is designed to facilitate Human-in-the-Loop (HitL) functionality, allowing for interactive user oversight during the scanning process. The specific Debian packaging files continue to be maintained in their dedicated repository at https://salsa.debian.org/J-mazz/sys-scan-debian_package. In support of the package's machine learning capabilities, I have also uploaded the model fine-tuning workflows and synthetic dataset curation scripts to https://salsa.debian.org/J-mazz/sys-scan-agent_MLops.git. The specific dataset utilized for this implementation has been published to the Hugging Face hub at https://huggingface.co/datasets/jmazz/sys-scan_synthetic_dataset_v2 to ensure transparency and reproducibility of the model's behavior. Thank you for your attention to these updates. Regards, Joseph Mazzini joseph@mazzlabs.works