cargo-auditable

Know the exact crate versions used to build your Rust executable. Audit binaries for known bugs or security vulnerabilities in production, at scale, with zero bookkeeping.

This works by embedding data about the dependency tree in JSON format into a dedicated linker section of the compiled executable.

Linux, Windows and Mac OS are officially supported. WebAssembly is also supported starting with v0.6.3. All other ELF targets should work, but are not tested on CI.

The end goal is to get Cargo itself to encode this information in binaries. There is an RFC for an implementation within Cargo, for which this project paves the way: rust-lang/rfcs#2801

Usage

# Install the tools
cargo install cargo-auditable cargo-audit
# Build your project with dependency lists embedded in the binaries
cargo auditable build --release
# Scan the binary for vulnerabilities
cargo audit bin target/release/your-project

cargo auditable works with any Cargo command. All arguments are passed to cargo as-is.

On nightly Rust

On nightly we can take advantage of Cargo's native SBOM precursor to record dependencies more accurately:

CARGO_BUILD_SBOM=true cargo +nightly auditable build -Z sbom --release

Due to a bug in Cargo you may have to touch src/* or cargo clean first if you also used cargo auditable without -Z sbom in the same project.

Through other tools

If you're not calling cargo directly and cannot change how it's invoked, you can use cargo auditable as a drop-in replacement for cargo. See here for details.

Adoption

Microsoft uses cargo auditable internally and previously maintained the data extraction library for Go.

Multiple Linux distributions build their Rust packages with cargo auditable: Alpine Linux, NixOS, openSUSE, Void Linux, Chimera Linux and Wolfi OS. If you install packages from their repositories, you can audit them!

Chainguard includes cargo auditable in their rust base container, with a default cargo wrapper to always call cargo auditable, so that Rust applications built using this container are auditable by default.

FAQ

Doesn't this bloat my binary?

In a word, no. The embedded dependency list uses under 4kB even on large dependency trees with 400+ entries. This typically translates to between 1/1000 and 1/10,000 of the size of the binary.

Can I make cargo always build with cargo auditable?

Yes! For example, on Linux/macOS/etc add this to your .bashrc:

alias cargo="cargo auditable"

If you're using a shell other than bash, or if using an alias is not an option, see here.

Is there any tooling to consume this data?

Vulnerability reporting

• cargo audit v0.17.3+ can detect this data in binaries and report on vulnerabilities. See here for details.
• trivy v0.31.0+ detects this data in binaries and reports on vulnerabilities. See the v0.31.0 release notes for an end-to-end example.
• grype v0.83.0+ detects this data in binaries and container images and reports on vulnerabilities.
• osv-scanner v2.0.1+ reads this data when scanning container images.

Recovering the dependency list

• syft v1.15.0+ has support for recovering this data and converting it to various formats. Older versions require the --catalogers all CLI option.
• docker supports embedding CycloneDX documents into container images. If you build a container image with docker buildx build --tag <namespace>/<image>:<version> --attest type=sbom --push . and use cargo auditable to build rust binaries in the Dockerfile, the SBOM attestation attached to the container image will include your rust dependencies. This is powerd by BuildKit Syft scanner.
• blint v2.1.3+ can recover this data and output it as CycloneDX.
• wasm-tools v1.227.0+ can recover this data from WebAssembly. Try wasm-tools metadata show.
• rust-audit-info recovers the dependency list from a binary and prints it in JSON.
• auditable2cdx recovers the dependency list from a binary and prints it in CycloneDX.

Can I read this data using a tool written in a different language?

Yes. The data format is designed for interoperability with alternative implementations. In fact, parsing it only takes 5 lines of Python. See here for documentation on parsing the data.

Besides that, Syft can read it and convert it to a multitude of formats. auditable2cdx can convert it to CycloneDX, which is understood by most tools. This conversion lets you feed this data even to tools you cannot modify.

What is the data format, exactly?

The data format is described by the JSON schema here. The JSON is Zlib-compressed and placed in a linker section named .dep-v0. You can find more info about parsing it here.

What about embedded platforms?

Embedded platforms where you cannot spare a byte should not add anything in the executable. Instead they should record the hash of every executable in a database and associate the hash with its Cargo.lock, compiler and LLVM version, build date, etc. This would make for an excellent Cargo wrapper or plugin. Since that can be done in a 5-line shell script, writing that tool is left as an exercise to the reader.

Does this impact reproducible builds?

The data format is specifically designed not to disrupt reproducible builds. It contains no timestamps, and the generated JSON is sorted to make sure it is identical between compilations. If anything, this helps with reproducible builds, since you know all the versions for a given binary now.

Does this disclose any sensitive information?

No. All URLs and file paths are redacted, but the crate names and versions are recorded as-is. At present panic messages already disclose all this info and more. Also, chances are that you're legally obligated have to disclose use of specific open-source crates anyway, since MIT and many other licenses require it.

What about recording the compiler version?

The compiler itself embeds it in v1.73 and later.

On older versions it's already there in the debug info. On Unix you can run strings your_executable | grep 'rustc version' to see it.

What about keeping track of versions of statically linked C libraries?

Good question. I don't think they are exposed in any reasonable way right now. Would be a great addition, but not required for the initial launch. We can add it later in a backwards-compatible way. Adopting the -src crate convention would make it happen naturally, and will have other benefits as well, so that's probably the best route.

Does this protect against supply chain attacks?

No. Use cargo-vet or cargo-crev for that.

Software Bills of Materials (SBOMs) do not prevent supply chain attacks. They cannot even be used to assess the impact of such an attack after it is discovered, because any malicious library worth its bytes will remove itself from the SBOM. This applies to nearly every language and build system, not just Rust and Cargo.

Do not rely on SBOMs when dealing with supply chain attacks!

What is blocking uplifting this into Cargo?

The RFC for this functionality in Cargo itself has been postponed by the Cargo team until the more foundational SBOM RFC.

That RFC has now been implemented and is available via an unstable feature. This opens the door to submitting an RFC for this functionality into cargo itself once again.