| name | weekly-review |
| description | Weekly CEO review posted every Monday. Summarizes the previous week's output (published articles, merged PRs, build health), reports cost per agent, identifies process debt, and proposes the next week's focus. Produces a Markdown attachment on the root company issue. |
Weekly Review
The weekly review is the CEO's strategic pulse check. It runs every Monday as part of the heartbeat cadence, before the regular morning brief.
What to Gather
Before writing:
- PRs merged to
AZANIR/artstroy master in the past 7 days (GitHub API)
- Articles published (slugs where
isDraft changed from true to false)
- Articles syndicated (Distribution Manager activity log)
- Build history: last 7 deploys — any failures? (GitHub Actions API)
- Paperclip agent activity: which agents were most active? (platform metrics)
- Open blockers older than 3 days
- Analytics Analyst's latest traffic report (if available in memory)
Output Format
Attach as Markdown to the root company issue (or post as a long comment):
## Weekly Review — Week of {YYYY-MM-DD}
### Published This Week
| Slug | Author | Published | Syndicated |
|---|---|---|---|
| {slug} | {author} | {date} | dev.to ✅ / ❌, Hashnode ✅ / ❌ |
### Pipeline Health
- PRs merged: {count}
- Build failures: {count} ({links if any})
- Open blockers > 3 days: {count} ({titles})
### Agent Activity (est. cost rank)
1. {most active agent} — {approximate token spend or issue count}
2. ...
(Flag any agent consuming >40% of total activity — may need scope review)
### Process Debt Identified
- [ ] {process that was done manually but should be automated}
- [ ] {recurring blocker pattern}
### Next Week's Focus
1. {priority 1 — specific, actionable}
2. {priority 2}
3. {priority 3}
### Strategic Notes
{1–3 sentences on anything worth tracking longer-term: traffic trends,
content gaps, infra concerns, team bottlenecks}
_Generated by CEO weekly-review at {datetime}_
Rules
- One per week: check if a weekly review for this week already exists before posting
- Cost vigilance: if one agent accounts for >40% of estimated spend without proportional output, flag it — could indicate a runaway loop or scope problem
- Process debt is a backlog item: for each item identified, create a real issue (don't just note it in the review and forget it)
- No blame, only patterns: focus on systemic process improvements, not individual agent failures
- After posting, proceed to the regular morning-brief for today
