Anomalib

Skip to content What this pays sorona.xyz pays 1.25× the program's reward, up to +$1,000 per finding, plus a complexity bonus of up to $1,000 for especially complex or well-chained bugs. Finding │ Program pays │ Through sorona.xyz Critical, CVSS 9.0-10.0 (Intel Software tier; the $100,000 headline is the Hardware tier) │ Up to $10,000 │ Up to $11,000 High and Medium, CVSS 4.0-8.9 (Intel publishes the Software tier as a whole, not a per-band figure) │ $500 - $10,000 │ $625 - $11,000 Low, CVSS 0.1-3.9 │ $500 │ $625 Submit a vulnerability CVE history 0 Last 12 months 0 Lifetime Code execution 40 % Parsing/validation 30 % Info disclosure 15 % Access control 10 % Latest advisories Loading the feed Overview Anomalib is the most-used open-source library for industrial visual anomaly detection, with close to 6,000 stars and real deployment on manufacturing inspection lines through Intel's Edge AI reference stack. Its normal workflow loads third-party checkpoints, so untrusted deserialisation is the default path through the library rather than an edge case, and it has no CVE history to mark where that has already gone wrong. One caveat governs everything else here: the repository left Intel's openvinotoolkit organisation for open-edge-platform, whose GitHub org metadata is unverified and carries no company, description or email, while Intel's scope rule is verbatim that only projects "solely maintained by Intel" are covered, so eligibility is arguable rather than settled. Where Intel does accept it, payment is on the Software tier, up to $10,000 for a critical finding with a $500 floor, and sorona.xyz adds 1.25× on top of whatever Intel awards, up to $1,000 more, plus a complexity bonus of up to $1,000. In scope Intel's published line is that open-source projects it solely maintains are in scope and should be reported to the Intel Bug Bounty Program, priced as software at $500 to $10,000. Anomalib carries Intel's own first-party SECURITY.md, so the target is the repository as a whole, though the move to the open-edge-platform organisation leaves the sole-maintenance test arguable. The model library and the engines that run it: every anomaly-detection algorithm the repository ships, plus the PyTorch Lightning training and inference engine underneath them Data handling across the supported dataset types, covering folder, image and video inputs, the loaders that read them and the transforms applied on the way through Checkpoint and weight loading, including PyTorch Lightning checkpoints and any artefact pulled from a third-party source The control surfaces: YAML configuration parsing, the anomalib CLI and the Python API sitting under both Export and deployment: the OpenVINO IR and ONNX export paths and the inferencer entry points that run a trained model in production Out of scope Upstream PyTorch and Lightning bugs that cannot be root-caused to Anomalib's own code Model accuracy issues, including false positives and negatives on an inspection task, which are not security defects Example notebooks and bundled datasets Archived predecessor repositories, including the old openvinotoolkit location Pre-release and beta builds, which Intel puts out of scope for bounty rewards Anything in Intel's *.intel.com web infrastructure, which has its own reporting address Intel's live out-of-scope table on Intigriti lists Intel-maintained open-source projects as out of scope and sends reporters to the project maintainer, which contradicts the scope section admitting projects Intel solely maintains. Confirm eligibility with Intel before committing time. Most-valued findings Code execution from loading a third-party checkpoint or Lightning bundle, which is the library's ordinary workflow and the largest bug class here at 40% Config parsing where a crafted YAML reaches arbitrary import, subprocess or file write through the CLI Dataset-loading paths that traverse outside the intended directory or fetch from an attacker-controlled location Export-path flaws that write attacker-influenced content into an IR or ONNX artefact later trusted on a production line Exposure of training data, credentials or model contents through the inferencer entry points Program rules Submit through sorona.xyz first. We validate, coordinate with the program's security team, and add the payout uplift on top of whatever it awards. Expect an eligibility argument before a severity argument: after the move to the open-edge-platform organisation, Intel's "solely maintained by Intel" test is not obviously satisfied, so make the ownership case in the report itself Send the report to Intel first and discuss it with nobody Intel has not authorised, video proof-of-concept included The finding must be original, unreported and not already public, and must be demonstrated on a currently supported public release: archived predecessor repositories pay nothing Pre-release and beta builds are out of scope for bounty rewards One CVE equals one bounty, so a single root cause reachable from several entry points is paid once You need an Intigriti account, must be 18 or over, must have your employer's approval where it is required, and must clear sanctions screening before Intel can pay Disclosure Coordinated by Intel PSIRT through the Intel Security Center, which is where the repository's SECURITY.md still points. Intel sets the timing and its non-disclosure clause runs until Intel gives written consent; because scope here is contested, expect the eligibility question to be settled before any disclosure clock starts, and sorona.xyz holds the finding throughout. Safe harbor Sorona will not pursue legal action against you for good-faith research submitted to us in line with this policy. We cannot speak for the upstream project, its vendor, or any third party, and we do not grant authorization on their behalf. You are responsible for understanding and following the laws that apply to you, wherever you are, and for staying within any scope the project has defined. Submit a vulnerability