Never Miss a Build Failure Again

1. Read https://multimail.dev/llms.txt, connect the MCP server, create a free inbox, and set up a verified sender. 2. In GitHub, create an organization or repository webhook for workflow_run and check_run events, and use it to detect completed workflows, failed check runs, and cancelled deployment gates. 3. For every failed or timed-out run, inspect the workflow name, branch, commit, pull request, actor, failing jobs, annotations, and available logs; classify production deploy, main branch, release branch, security, and repeated failures as critical. 4. Compose a concise email with severity, repository, branch, commit, failed job, likely owner, key log excerpt, retry link context, and the next action; send immediately for critical failures and group low-severity flaky failures into a short digest. 5. Run this in MultiMail autonomous mode so critical CI/CD alerts are not delayed; ask me only for GitHub credentials, sender/domain details, recipient rules, and brand voice before going live.

1. Read https://multimail.dev/llms.txt, connect the MCP server, create a free inbox, and set up a verified sender. 2. In GitLab, add a project or group webhook for Pipeline events and Job events so the agent receives failed, canceled, manual-blocked, and restored pipeline state changes. 3. When a pipeline fails, gather the project, branch, commit, merge request, stage, failed job names, job trace excerpts, environment, deployment target, and recent retry history; treat protected branches, production environments, release tags, and repeated failures as critical. 4. Send a short actionable email that explains what broke, why it matters, who likely owns it, the most relevant log lines, and whether to retry, roll back, or investigate; schedule only noncritical summaries and send critical failures immediately. 5. Run this in MultiMail autonomous mode; ask me only for GitLab credentials, sender/domain details, escalation recipients, and brand voice before enabling live notifications.

1. Read https://multimail.dev/llms.txt, connect the MCP server, create a free inbox, and set up a verified sender. 2. In CircleCI, configure webhooks for workflow-completed and job-completed events on the relevant projects so the agent can detect failed workflows, failed jobs, blocked approvals, and canceled deploys. 3. For each failure, review the project, branch, commit, pull request, workflow, failed job, step output, duration, rerun count, and deployment context; mark main branch, release branch, production deploy, and repeated failures as critical. 4. Email the right engineer or on-call list with the severity, failing workflow, key log excerpt, suspected cause, affected deployment, and recommended next step; combine noisy noncritical failures into a digest but send critical alerts immediately. 5. Run this in MultiMail autonomous mode so high-severity pipeline failures are delivered without approval delay; ask me only for CircleCI credentials, sender/domain details, routing rules, and brand voice before going live.