GLIMPS Malware for DevSecOps
Automatically analyze every artifact in your CI/CD pipelines to detect malicious components before production deployment.
Secure your software supply chain at every stage
Protecting Against Compromised Open-Source Dependencies
According to ReversingLabs’ 2026 Software Supply Chain Security Report, malicious open-source packages have increased by 73%.
Detect before deployment
Every binary, library, and resource is analyzed by more than 25 detection engines, including around ten AI-based engines, before it reaches production environments.
~3-second average verdict
Analysis does not slow down release cycles. Each artifact is qualified in seconds, compatible with DevOps speed constraints.
How does GLIMPS protect your CI/CD pipelines?
GLIMPS Malware for DevSecOps integrates into continuous integration and deployment pipelines to analyze each component at different stages of the development lifecycle:
GLIMPS Malware REST API – Call the API directly from pipeline scripts (Jenkins, GitLab CI, GitHub Actions, or any other tool). At each critical step (dependency import, build, pre-deploy), files are submitted for analysis. The verdict is returned as JSON with the detection score and identified malware families. Based on the result, the pipeline continues or automatically stops.
Open source Python and Go libraries – py-gdetect and go-gdetect make integration easier and speed up development of custom scanning steps.
Host Connector for third-party deliverables – As part of supply chain workflows, Host Connector can monitor an intake folder for deliverables provided by external vendors (binaries, ISO, OVA, archives). Each dropped file is automatically analyzed and quarantined if a threat is detected before it is integrated into the build chain.
All results can be reviewed in GLIMPS Malware Expert (optional) for deep investigation, triage, and retrohunt.
Analysis capabilities built for DevSecOps
Multi-script LLM: local, sovereign semantic analysis
Our proprietary LLM semantically analyzes JavaScript, Python, and PowerShell code to detect malicious behavior hidden in imported scripts and dependencies. Unlike signature-based approaches, this analysis understands code intent and identifies threats even when heavily obfuscated. The LLM runs locally with no external connection: your source code and artifacts never leave your environment.
Multi-engine binary detection
More than 25 detection engines, including around ten AI-based engines (Deep Learning, Machine Learning), complemented by antivirus engines and your own YARA rules, analyze each compiled binary and library to identify backdoors, trojans, and supply chain attacks.
Automatic archive extraction
Submitted archives (ZIP, RAR, 7Z…) are automatically extracted and analyzed by GLIMPS Malware. Every contained file is scanned by the detection engines.
Cache and performance
Built-in caching prevents resubmission of already analyzed artifacts. Designed for high pipeline volumes (hundreds of thousands of artifacts per month) without impacting build times.
GLIMPS Malware for DevSecOps: who is it for?
Context: You embed security into CI/CD pipelines and must ensure every deployed component is free of malicious code.
What GLIMPS delivers: Ready-to-use API and open source libraries to add a security scan at each stage of your pipeline. The multi-script LLM detects threats in imported JavaScript, Python, and PowerShell code, going beyond traditional SCA tools. The LLM is local: your code never goes to an external service.
Context: You use open source libraries and third-party components. A compromised dependency can compromise the entire application.
What GLIMPS delivers: Automatic, transparent scanning of every imported dependency. No manual action required: analysis is integrated into the pipeline and the build automatically stops on detection.
Context: Supply chain attacks are increasing and targeting open source components. Your organization’s software attack surface is hard to control.
What GLIMPS delivers: A sovereign AI detection layer that protects developments against backdoors, compromised libraries, and malicious binaries. Aligned with GDPR and NIS2. Every vendor deliverable is analyzed before it enters your build chain.
How customers use GLIMPS Malware for DevSecOps
Automated scanning of open source dependencies
A development team regularly imports libraries from public repositories (npm, PyPI, GitHub). GLIMPS analyzes each new dependency in the CI/CD pipeline. The multi-script LLM examines JavaScript and Python code to detect malicious behavior. If a threat is identified, the build is blocked before deployment.
Verifying third-party deliverables (supply chain)
An organization receives modules developed by external vendors: binaries, ISO, OVA, archives. Host Connector monitors the intake folder and automatically analyzes each deliverable. Compromised files are quarantined before they enter the build chain, protecting against supply chain attacks.
Final artifact control before production
Before each deployment, the pipeline submits the final deliverable to the GLIMPS Malware API for a last check. The verdict confirms the absence of malicious code in the produced binary, ensuring a reliable, secure release.
Why choose GLIMPS Malware for DevSecOps?
Multi-script LLM: a unique, expanding capability
GLIMPS is the only vendor providing semantic analysis of JavaScript, Python, and PowerShell using a proprietary LLM. Threats hidden in scripts and dependencies are detected even when heavily obfuscated. More languages will be supported soon (BAT, Rust, Go).
Proprietary AI detection
GLIMPS’ code conceptualization technology, developed within the French Ministry of the Armed Forces by GLIMPS’ founders, identifies backdoors, variants, and zero-day threats in compiled binaries. Detection rate above 99%.
Used by the French Armed Forces
Used by the French Armed Forces and incubated by the French Ministry of the Armed Forces. French Tech 2030 award winner.
Frequently asked questions
Integration is done through the GLIMPS Malware REST API, called directly from pipeline scripts (Jenkins, GitLab CI, GitHub Actions, or any other tool). Open source Python (py-gdetect) and Go (go-gdetect) libraries are available on GitHub to speed up integration. You can be operational within a few hours.
GLIMPS analyzes all file types: compiled binaries, scripts (JavaScript, Python, PowerShell), libraries, archives (ZIP, RAR, 7Z…), disk images, and containers. The multi-script LLM adds dedicated semantic analysis for interpreted code.
Analysis takes about ~3 seconds on average per artifact. A cache prevents resubmission of already analyzed artifacts. The solution is designed to handle large volumes without impacting build times.
Yes, GLIMPS Malware for DevSecOps is available as SaaS (hosted in France), on-premise, and in SecNumCloud environments. All AI models, including the multi-script LLM, run locally and do not require any external connection. Your source code never leaves your environment, regardless of deployment mode.
SCA tools identify known vulnerabilities in open source dependencies using CVE databases. GLIMPS goes further by analyzing the actual content of files with AI: detecting backdoors, obfuscated malicious code, unknown variants, and malicious scripts via the multi-script LLM. The approaches are complementary.
Request a demo
Are you interested in our products? Contact our experts and discover the full power of our solution!
Request a demo
