/drain

Drain the Ready column autonomously — the overnight workhorse that loops /work-ticket → /review-pr → agent-merge → post-merge validation for each safely-classified ticket while you sleep.

When to use it

Use /drain when the Ready column is groomed, every ticket has a safety:* label, and you want the framework to advance multiple tickets through the full implementation cycle unattended. The drain run pre-flights production health, topo-sorts dependencies, filters by safety class, and works through the queue one ticket at a time. Each ticket gets its own PR, CI run, reviewer verdict, merge gate, and production validation before the next ticket starts.

/drain is specifically the merge loop — it moves code from Ready to Done. For promoting a safety:flagged feature from “dark on production” to “visible to users,” reach for /goal.

The operator’s machine must stay on and the Claude Code session must stay open (v1 trade-off — remote unattended operation is a future ticket). You make the flag-flip release decision in the morning; drain only ships code, never lifts a feature flag.

Safety classes accepted by default

How it fits

What it does (quick)

• Pre-flights production health (Sentry baseline + board access + agent-merge.yml presence)
• Snapshots the Ready column; topo-sorts by Depends on: lines; filters by safety class
• Emits a plan summary as a comment on the originating drain invocation
• Loops one ticket at a time: /work-ticket → CI watch → /review-pr → bridge dispatch → Render deploy poll → post-deploy validation
• Writes state to disk at 9 trigger points so /drain --resume can re-enter after an API interruption
• Emits an audit comment on each ticket naming the merge commit, deploy ID, and validation result
• Posts a wake-up summary at end of run via scripts/emit-drain-summary.mjs

Invocation shapes

/drain                          # process safety:flagged + safety:internal; no end time
/drain --until 06:00            # stop at 6 am if Ready is not empty by then
/drain --include-reversible     # opt in to safety:reversible tickets (rate-limited)
/drain --max-tickets 5          # cap the number of tickets processed
/drain --dry-run                # emit the plan summary; do not execute
/drain --resume <drainId>       # resume an interrupted drain run
/drain --resume                 # auto-detect the most recent interrupted run

Related commands

• /groom-backlog — fills Ready before invoking /drain
• /work-ticket — the per-ticket workhorse drain calls internally
• /review-pr — auto-chained from /work-ticket on green CI
• /goal — picks up from where /drain leaves off; promotes a dark-shipped feature flag to GA

Canonical spec: .claude/commands/drain.md