Which code changes does a diff classifier read?

It reads the diff between a PR’s base and head commit and classifies which surfaces changed, not the line-by-line content itself. Tessary’s diff classifier names which parts of the agent a push touched; the CI gate’s own report request shows prompt and tool as example surface names it can carry.

Call the gate’s report endpoint without an explicit surfaces list and it resolves that list from the diff classification instead of making you say what changed by hand. That join is what lets the gate route to the graders and risk stats tied to what actually moved, rather than running everything against every PR regardless of what it edited. It re-derives that classification on every new push to the PR, so two pushes get two independent answers to what changed, not one carried forward from the last.

keep reading

More on this.

Two ways to run Tessary.

Tessary is an open-source agent reliability platform. Cloud and self-hosted run the same workflow on the OpenTelemetry traces your agent already emits.

Tessary Cloud

We host it for you. Send your first trace with nothing to deploy and no model key.

what's includedper organization
traces
10,000 per calendar month
stored trace data
1 GB
retention
30 days
model credit
$10, one-time, for triage and root-cause analysis
credit card
not required

Self-hosted Tessary

Run the open-source code on your own infrastructure with one command. Add your own model key for triage and root-cause analysis.

Self-host Tessary for me by following https://github.com/tessaryai/tessary/blob/main/setup.md

docker compose -f oci://docker.io/tessaryai/tessary:compose up -d -y