Review release readiness in Rootly with AI
Assess a proposed release against the team’s policy, test evidence, current incidents and rollback plan. Identify missing checks before a human makes the release decision.
Review release readiness in Rootly with AI when a small change might still affect a busy service or shared dependency. The deploy-check skill considers the proposed artefact, target environment, testing, human review and rollout safeguards. It also checks for an incident or change freeze that could make a technically sound release badly timed.
For a checkout feature behind a flag, the relevant questions include what was tested, where the feature will run and whether a concurrent infrastructure change affects the same path. A missing test result remains missing. The output gives a scoped recommendation and states what would change it.
What it does
- Compare a change with the customer’s own release rules.
- Review open or monitoring incidents on affected services.
- Identify concurrent upgrades of shared dependencies.
- Check a canary, feature flag or rollback safeguard.
- Separate verified approvals from outstanding evidence.
Install and use
- Install the skill, provide the proposed change and release policy, and connect authorised Rootly data through its official read-only MCP. Supply permitted repository or CI evidence where needed.
- Ask, “Review the staged checkout release. Check its affected dependencies and explain which remaining checks would change the verdict.”
- Verify the target environment and evidence timestamps with the release owner before acting.
FAQ
Source
Made by Rootly. Original skill: deploy-check, published on GitHub under the Apache-2.0 licence.
View on GitHub