Inspektor Gadget: Results from the first security audit
Inspektor Gadget, the open source eBPF-based toolkit for Kubernetes observability, has completed its first independent security audit coordinated by OSTIF and executed by Shielder. The audit found three vulnerabilities (two Medium, one Low) with no Critical or High severity issues, and delivered six hardening recommendations. All reported issues have been patched in releases v0.48.1 through v0.50.1, and the project received a positive assessment of its overall security posture.
Inspektor Gadget Passes First Independent Security Audit With Strong Results
The eBPF-based toolkit for Kubernetes observability and Linux host inspection, Inspektor Gadget, has completed its first independent security audit — and the findings paint a reassuring picture for the thousands of teams relying on it in production. Coordinated by the Open Source Technology Improvement Fund (OSTIF), funded by the CNCF, and executed by the security firm Shielder, the audit examined the project's codebase, deployment patterns, and overall security posture with rigor.
The result? Three vulnerabilities discovered — two rated Medium severity and one Low — with no Critical or High findings. All reported issues have been patched in subsequent releases.
What Is Inspektor Gadget and Why Does It Need Auditing?
Inspektor Gadget is a framework that uses eBPF (extended Berkeley Packet Filter) to collect and inspect data on Kubernetes clusters and Linux hosts. It packages, deploys, and manages "gadgets" — eBPF programs distributed as standard OCI container images. This means operators can load kernel-level tracing programs without rebuilding container images, injecting sidecars, or restarting workloads.
Key capabilities include:
- OCI-packaged gadgets: eBPF programs distributed as standard container images, compatible with any registry
- Zero-instrumentation deployment: No rebuilds, sidecars, or debuggers needed
- Multiple deployment modes: Local host, remote daemon, or full Kubernetes integration
Because Inspektor Gadget runs with elevated privileges on shared infrastructure, an independent security audit was a natural — and necessary — step as the project continues to see growing adoption across cloud native environments.
Audit Scope and Methodology
OSTIF engaged Shielder in early 2026, with two security researchers performing the assessment using a multi-layered approach:
- Collaborative threat modeling with Inspektor Gadget maintainers
- Manual source code review across the full codebase
- Dynamic testing on dedicated lab environments reflecting production deployments
- Static analysis using Semgrep and GoSec
- AI-assisted code review for broader coverage
The researchers built three test environments mirroring real-world Inspektor Gadget usage: a local Linux host deployment, a remote daemon setup, and a full Kubernetes deployment running on minikube.
The Vulnerabilities: Two Medium, One Low
Medium — Command Injection in ig image build (CVE-2026-24905)
The image build process used Makefiles that embedded user-controlled input without proper escaping, creating a command injection vector. This is most dangerous in CI/CD pipelines building untrusted gadgets. Fixed in v0.48.1.
Medium — Denial of Service via Event Flooding
A malicious container could flood the 256 KB eBPF ring buffer, causing silent event drops from other containers. For organizations relying on Inspektor Gadget for security monitoring, this could mask malicious activity. Fixed in v0.50.1.
Low — Unsanitized ANSI Escape Sequences (CVE-2026-25996)
Terminal output rendered events without sanitizing ANSI escape sequences, allowing a compromised container to inject escape codes into an operator's display. Fixed in v0.49.1.
Hardening Recommendations for the Future
Beyond the specific vulnerabilities, Shielder delivered six hardening recommendations to reduce the project's attack surface:
- Enforce TLS by default on TCP listeners instead of requiring an explicit flag
- Pin and verify external dependencies in CI/CD with hash or signature verification
- Implement a Kubernetes namespace blocklist to prevent tracing on sensitive namespaces like
kube-system - Restrict remote daemon access to prevent host-level tracing from compromised client connections
- Automate third-party vulnerability scanning for project dependencies
- Reduce RBAC permissions on the DaemonSet pod, particularly the
nodes/proxyGET permission
The maintainers are systematically working through these recommendations. Several have already been addressed, while the RBAC refactor and namespace blocklist remain in progress.
Gadget Bypass Testing: The Cat-and-Mouse Reality of eBPF Security
One of the most technically revealing audit components was the gadget bypass testing. The researchers asked a crucial question: could a compromised container perform operations a gadget is meant to trace without triggering any events? They identified six bypass scenarios, including:
- Using newer Linux syscalls that certain gadgets don't hook (e.g.,
openat2instead ofopenat) - Evasion through
io_uring— Linux's asynchronous I/O interface - Statically linked libraries that bypass dynamic tracing hooks
These findings highlight the inherent challenges of kernel-level tracing. As Linux evolves with new syscalls and subsystems, eBPF-based tools must keep pace. The Inspektor Gadget team has already addressed several gaps and is documenting inherent limitations so operators understand what eBPF tracing can and cannot guarantee under adversarial conditions.
What Teams Should Do Now
Organizations running Inspektor Gadget should update to v0.50.1 or later, which includes fixes for all three reported vulnerabilities. Shielder's final assessment was that "the overall security posture of Inspektor Gadget is adequately mature from both a secure coding and design point of view."
For the broader cloud native community, this audit demonstrates the open source security ecosystem functioning as intended: a project reaching sufficient adoption to warrant independent review, OSTIF coordinating a qualified engagement, researchers performing transparent work, maintainers shipping fixes, and the full report published for informed decision-making.
The complete audit findings — including the full PDF report and individual CVE details — are available on the Inspektor Gadget blog and the OSTIF website.



