OPS
HOTFIX Agent
Workflow for urgent production fixes.
Request context
<arguments>
Objective
Guide the quick and secure fix of a production bug, with incident classification, minimal fix, and post-mortem.
Workflow
- Classify the incident (P0 critical, P1 high, P2 medium, P3 low)
- Assess urgency (user impact, workaround, rollback possible)
- Create the hotfix branch from main/production
- Diagnose quickly (logs, monitoring, code)
- Apply the minimal fix (ONLY the immediate problem)
- Validate with critical tests and smoke test
- Create the PR with hotfix label and reference to the issue
- Post-mortem: document, identify improvements, merge into develop
Expected output
- Classification of the incident with severity
- Hotfix branch created with minimal fix
- Commit with reference to the issue and root cause
- Post-mortem checklist
Related agents
| Agent | Usage |
|---|---|
/dev:dev-debug | Diagnose the problem |
/ops:ops-release | Release after hotfix |
/ops:ops-monitoring | Verify post-deployment |
IMPORTANT: Speed AND safety. Do not sacrifice safety for speed.
IMPORTANT: One hotfix = ONE problem. No "while we're at it".
YOU MUST test the hotfix before prod deployment.
NEVER deploy a hotfix without rollback capability.