the file that tells Yocto how to fetch and build one piece of software · the vendor and product names a recipe is filed under in vulnerability databases
What it is
gd
Notes on this component
libgd (the GD graphics library) creates and reads images in PNG, JPEG, GIF, BMP and other formats, which web pages and tools on a device use to draw charts or resize pictures. The project's release notes list the security fixes with their CVE identifiers, and a file-format parser works on input from outside. The recipe's CVE_PRODUCT is libgd with no vendor, so a lookup matches on the product name libgd alone, not on the recipe name gd.
Where it hides in a firmware SBOM
Yocto lists it by recipe name. Vulnerability databases file the same code under a vendor:product pair taken from the recipe's CVE_PRODUCT, so the two names often differ.
The identifiers we assign
the product name vulnerability databases use to say what a flaw affects
cpe:2.3:a:*:libgd:{version}:*:*:*:*:*:*:*
{version} is filled in from your build. The version is the recipe version in your image.
The SBOM check adds these identifiers to a component your build listed without them.
Honest gaps
Vendor: not recorded. The recipe names the product but not who publishes it, and we do not guess vendors. A * vendor matches records for any product called libgd, including unrelated ones.
Vulnerability data changes daily. Check the version you ship in the SBOM check.
CRA notes
List this component in your SBOM. Annex I, Part II (1): Know what is in your product. Identify and document its components and their vulnerabilities, including by producing a software bill of materials (SBOM) in a common machine-readable format that covers at least the top-level dependencies.
Ship the version with no known exploitable vulnerabilities at release. Annex I, Part I (2(a)): When you place a release on the market, it must not ship with known vulnerabilities that can be exploited in your product. "Exploitable" means an attacker could effectively use it under practical operating conditions (Article 3(41)), not only in theory or in a lab. The Commission's draft guidance (C(2026) 5252, points 231 to 237; a draft, not binding) gives three ways a vulnerability is "known". One: it is listed in a relevant public vulnerability database, such as the European vulnerability database (EUVD) or another prominent one. Two: you learned of it from information that is not public, for example a security researcher's coordinated disclosure or your own testing and analysis; a vulnerability you know about and have not disclosed still counts. Three: it has been publicly and prominently reported in reliable media, whether specialist security publications or general news. A report is not yet proof: it does not by itself mean the vulnerability can be exploited in practice or applies to your product, so investigate promptly and confirm whether it does. A limited time may pass between the first report and that confirmation. If a vulnerability becomes known shortly before release, you decide from your cybersecurity risk assessment whether the release can go out or the vulnerability must be fixed first. Weigh its severity, exploitability and potential impact, and the risks to the product once in use; you may also weigh the risks of delaying the release, for example because it fixes other exploitable vulnerabilities. After release, the Part II vulnerability handling requirements still apply. Check your components against vulnerability databases before each release, and fix or justify what you find.
Source: Annex I of Regulation (EU) 2024/2847 (Cyber Resilience Act), legal state 2026.10.13-1.