content-calendar-sheets

Every Monday morning, reads a ContentPlan Sheet and sends a digest of the upcoming week's content — platform, topic, draft snippet, status, assignee — to the configured contact. Anyone on your team who needs to know "what are we posting this week" gets the answer over coffee, without opening a tool.

Why this exists

The original content-calendar hub workflow reads a local CSV but sends a static message ("Here's your content plan for this week. Time to create!") with no actual plan content in the body. You receive a Monday reminder telling you to go look at the plan — not the plan itself.

This variant reads the actual entries from a Sheet you edit on your phone, formats them into the digest, and groups by day. The Sheet doubles as your editorial workflow (status: planned → drafted → ready → posted), and the digest surfaces what's ready to publish vs what still needs work.

Setup

1. Credentials (Credentials view): - google-workspace — Google Workspace service-account JSON
2. Runtime environment (Runtime Environments view): - python — auto-created on every install
3. ContentPlan Sheet (one-time): - Create a Sheet with a tab ContentPlan and 8 columns: A date | B platform | C topic | D draft E hashtags | F status | G assignee | H notes - Row 1 is the header. Data starts at row 2. - Maintain it as your editorial backlog. Add far-future entries; this workflow only surfaces the current week. - status values: planned / drafted / ready / posted / skipped. The workflow doesn't update status — that's a human decision. The digest just rolls up the counts.
4. Edit workflow.yaml — replace the two placeholders: - <<CONTENT_PLAN_SHEET_ID>> — spreadsheet ID - <<OWNER_PUBLISHER>> — user name for your contact
5. Schedule every Monday at 7:00 AM (or your team's preferred start time).

Sample digest

CONTENT CALENDAR — week of 2026-05-11 to 2026-05-17

Monday May 11:
  [ready]     blog       Why we switched from Postgres to ClickHouse   (owner: Alice)
      | We migrated 400GB last quarter and learned a few things — most…
      | tags: #engineering #databases #postgres

Tuesday May 12:
  [drafted]   twitter    Thread: 5 things people get wrong about CDC
      | 1/ Change Data Capture is not just a backup strategy. …
      | tags: #data #infra

Wednesday May 13:
  [planned]   linkedin   Hiring announcement: Senior Backend           (owner: Bob)

Friday May 15:
  [ready]     newsletter May product update — index rebalancing       (owner: Carol)
      | Carol's draft is ready for review; link to doc in the Sheet
      | note: needs design review before send

Status this week: ready=2, drafted=1, planned=1

2 entries marked READY — review and publish them today.

Pipeline

1. compose — runtimeenv (python). Reads ContentPlan!A2:H, filters to the current week (Monday-Sunday), groups by date, formats each entry with status / platform / topic / draft snippet / hashtags / notes, computes status rollup and a "ready to publish" callout. Prints to stdout.
2. notify — message. Pipes the digest to your owner user.

Customization

• Week start — WEEK_START_DAY = 0 is Monday. Change to 6 for Sunday-Saturday weeks (US convention).
• Draft snippet length — change e["draft"][:180]. Trim shorter if your subscribers are SMS-only.
• Add to-do reminders — extend the script: anything still planned on Friday triggers a separate alert.
• Per-platform digest — if your team has separate owners per platform, split the digest by platform and send a different message task to each platform owner.
• Auto-publish hook — add a posted status transition path: a second workflow runs daily, finds ready rows for today, posts via the platform's API (Twitter/LinkedIn via Workspace --attach? Buffer? Hootsuite? — depends on your channel), and updates status to posted. Out of scope for this digest workflow but a natural follow-on.

Differences from the original content-calendar

Verified

gog sheets get --json --results-only syntax verified against the bundled gog v0.19.0 (2026-06-01) — returns nested arrays of strings (re-verify on first run); row width padded to 8 columns to handle ragged data. Python date math standard datetime.timedelta.

Source

See workflow.yaml.