Go implementation of the in-toto specification.
To read the documentation along with some examples, run:
godoc -http :8080
and navigate to localhost:8080/pkg/github.com/in-toto/in-toto-golang/
Alternatively, you can use pkg.go.dev.
A very simple example, just to help you starting:
package main
import (
"time"
toto "github.com/in-toto/in-toto-golang/in_toto"
)
func main() {
t := time.Now()
t = t.Add(30 * 24 * time.Hour)
var keys = make(map[string]toto.Key)
var metablock = toto.Metablock{
Signed: toto.Layout{
Type: "layout",
Expires: t.Format("2006-01-02T15:04:05Z"),
Steps: []toto.Step{},
Inspect: []toto.Inspection{},
Keys: keys,
},
}
var key toto.Key
key.LoadKey("keys/alice", "rsassa-pss-sha256", []string{"sha256", "sha512"})
metablock.Sign(key)
metablock.Dump("root.layout")
}
Download the source, run make build
.
The CLI reference can be found in the autogenerated docs.
This implementation of in-toto has been integrated with SPIFFE/SPIRE. The integration is made possible by ITE-7, an enhancement that adds support for X.509 signing to in-toto.
To run the demo, pull down the source code, install Go, and run
make test-verify
. This will use openssl to generate a certificate chain.
To run the demo using SPIRE, pull down the source code, install Go and Docker,
and run make test-spiffe-verify
.
SPIFFE compliant Leaf certificates are generated with SVIDs corresponding to functionaries. These certificates are consumed by in-toto to sign link metadata and the layout policy.
During the in-toto verification process, certificate constraints are checked to ensure the build step link meta-data was signed with the correct SVID.
Currently the following constraints supported:
{
"cert_constraints": [{
"common_name": "write-code.example.com",
"dns_names": [
""
],
"emails": [
""
],
"organizations": [
"*"
],
"roots": [
"*"
],
"uris": [
"spiffe:https://example.com/write-code"
]
}, {
"uris": [],
"common_names": ["Some User"]
}]
}
This golang implementation was focused on verification on admission controllers and kubectl plugins. As such, it focused on providing a strong, auditable set of core functions rather than a broad and (possibly) unstable feature set. In other words, we believe that the current feature set is stable enough for production use.
If any of these features are necessary for your use case please let us know and we will try to provide them as soon as possible. Alternatively we welcome pull requests with feature additions!