Problem Detection and Resolution
Make problems visible early
Agile risk management is continuous, not a document written once. Maintain a risk-adjusted backlog, and handle each risk with ROAM - Resolved, Owned, Accepted, or Mitigated - naming an owner. Track a risk burndown so exposure trends down. Impediments go on a visible impediment log and are escalated fast; a blocker raised at the retrospective has already cost the team an iteration.
Diagnose before you fix
Use 5 Whys and fishbone diagrams to reach root cause rather than symptom. Pareto analysis focuses effort on the few causes producing most defects. Control charts separate common-cause variation (the system) from special-cause spikes (an event) - reacting to common-cause noise makes things worse. Track cycle time, lead time, throughput and WIP; Little's Law ties them together. WIP limits deliberately expose bottlenecks instead of hiding them behind queues. Adopt a stop-the-line mentality: fix a defect where it is found rather than passing it downstream, and treat escaped defects and technical debt as signals about the process, not just the code.
Traps
A risk register with no owners or actions; waiting for the retrospective to escalate; blaming individuals instead of examining the system; adding process to fix a trust problem; declaring a defect fixed when it is coded but untested.
Rule of thumb: surface it immediately, swarm to protect flow, find the root cause, then put the fix in the backlog with an owner and a date.