incident

# zuvo:incident — Incident Response & Postmortem A structured framework for production incident investigation. Builds a timeline from git history, CI/CD deploys, and error tracking, identifies suspect commits, assesses impact, recommends remediation, and generates a blameless postmortem document. **Scope:** Production incidents where something is broken, degraded, or behaving unexpectedly. Investigation, root cause analysis, and postmortem generation. **Out of scope:** Actually applying fixes (use `zuvo:debug` or `zuvo:build`), code quality sweeps (`zuvo:code-audit`), performance investigation without an active incident (`zuvo:performance-audit`). ## Argument Parsing Parse `$ARGUMENTS` for these flags: | Flag | Effect | |------|--------| | `[description]` | What happened (free text) | | `--since [time]` | When the incident started (ISO-8601, relative like `2h`, or natural language like `yesterday 3pm`). Default: `24h` ago | | `--service [name]` | Affected service, endpoint, or component | | `--sev [1-4]` | Override auto-detected severity (1=critical, 4=low) | | `--revert` | Include revert recommendation with exact command | | `--comms` | Generate communication templates (internal + customer-facing) | | `--dry-run` | Analyze only, do not create postmortem file | Flags can be combined: `zuvo:incident payments returning 500 --since 2h --service /api/payments --sev 1 --revert --comms` --- ## Environment Compatibility Read `../../shared/includes/env-compat.md` for agent d...