Release Governance and Readiness

A release is a decision about evidence, not just a successful build. Teams need to know which checks ran, which exceptions were accepted, and who approved the risk before a change moves forward.

Make readiness an explicit workflow

A Farna Control pipeline can collect inputs from your repository, issue tracker, CI system, and other configured connectors. It can validate a readiness record, run agents against the evidence, apply evaluation gates, and pause for human approval before the next stage. Each run records the inputs, outputs, gate decisions, and final artifact.

The checks themselves come from the tools and rules your team configures. Farna Control does not claim that a release is safe because a pipeline completed. It records the evidence and the decisions that your process produced.

Keep the approval boundary visible

A human-in-the-loop gate can require review before a release action. A human-only gate keeps the decision with a human principal. The reviewer receives the artifact and upstream context, and the decision becomes part of the append-only audit chain.

You can set autonomy per pipeline or node. Keep approval for release classes that need it, then use evaluation results and observed runs as evidence when deciding whether a lower-friction path is appropriate.

Current and planned connections

Repository reads and writes, issue tracking, webhooks, agent signals, and run evidence are available through the connectors and triggers you configure. CI-runner connectors can read statuses and logs today; dispatching a pipeline node to start a CI-runner job is still in development. Deployment actions remain dependent on the deployment connector or workflow you provide.

Where to go next

• Audit Trail for AI-Generated Code: the record created for each run
• Human-in-the-Loop: review gates and decisions
• Existing Agent Runtimes: connecting the systems that run your work