intake

Intake — GitHub Issues → Beans Epics

Import open GitHub Issues as beans epics so the orchestrator has a single work queue.

Prerequisites

• gh CLI authenticated (gh auth status)
• beans CLI available

How it works

1. Fetch open GitHub Issues that haven't been imported yet.
2. For each issue, create a beans epic with the issue body as the spec.
3. Embed a <!-- gh-issue: <number> --> marker in the epic body for dedup.

Steps

1. Fetch open issues

gh issue list --state open --json number,title,body,labels --limit 100

2. Check for already-imported issues

For each issue, check if a bean with a matching marker already exists. The marker convention is <!-- gh-issue: <number> --> embedded in the epic body.

beans list --type epic --json | jq -r '.[].body' | grep -q '<!-- gh-issue: 42 -->'

If a bean already contains this marker, skip it.

3. Create beans epic

For each new issue:

beans create "<issue title>" \
  --type epic \
  --body "<!-- gh-issue: <number> -->

<issue body>"

The marker must be the first line of the body so it's easy to find.

4. Report

List what was imported:

echo "Imported: #<number> → <bean-id>"

Mapping

GitHub Issue field	Beans epic field
number	body marker <!-- gh-issue: <number> -->
title	title
body	body (after marker line)
labels	not mapped (could be priority later)

Notes

• This is a one-way sync: GH Issues → beans. Beans don't write back to GH Issues.
• PR linking happens at finalize time (docs/workflow.md step 3), not during intake.
• Re-running intake is safe — duplicates are skipped via body marker.
• Closed GH Issues are not imported. If an issue is closed after import, the orchestrator handles it by checking bean status, not GH state.