github:last-week-org

Deep weekly analysis across all rossoctl org repositories. Active repos get full analysis; quiet repos get a one-liner summary.

Variables

Set at session start:

export OWNER=<org-or-user>
export REPO=<repo-for-report-issue>   # where the weekly report issue is posted

When to Use

• Org-wide weekly standup / leadership update
• Cross-repo coordination check
• Tracking health of the entire rossoctl org

Auto-approved: All gh read commands and gh repo list are auto-approved.

Workflow

Phase 0: Gather Org Data (MUST run first)

Run the org data gathering script to snapshot all repos:

./.github/scripts/reports/weekly-report-org-data.sh 7 > /tmp/rossoctl/github/org-data-gather.log 2>&1; echo "EXIT:$?"

Then read the org summary (small file, OK inline):

jq . /tmp/rossoctl/github/data/org-summary.json

This creates:

• /tmp/rossoctl/github/data/<repo>/ — Per-repo JSON files (merged-prs, open-prs, open-issues, new-issues, ci-runs)
• /tmp/rossoctl/github/data/org-summary.json — Totals and per-repo counts with active/quiet classification

IMPORTANT: All subsequent phases MUST read from these JSON files, NOT re-query gh. This ensures consistency.

Phase 1: Org-Wide Summary Table

Read totals from org-summary.json and present:

## Org-Wide Summary

| Repo | Merged PRs | Open PRs | Open Issues | New Issues | CI Pass Rate | Status |
|------|-----------|----------|-------------|------------|-------------|--------|
| rossoctl | ... | ... | ... | ... | .../... | active |
| repo2 | ... | ... | ... | ... | .../... | quiet |
| **TOTAL** | **N** | **N** | **N** | **N** | | |

Order repos by activity (most merged PRs + new issues first).

Linking Convention (ALL phases)

IMPORTANT: All issue and PR references in the report MUST use absolute GitHub URLs, not shorthand references like #N or rossoctl/<repo>#N. Shorthand references are ambiguous when the report is posted as a GitHub issue — GitHub auto-links them relative to the repo where the issue lives.

Use this format for every repo, including rossoctl itself:

• Issues: [rossoctl/<repo>#N](https://github.com/rossoctl/<repo>/issues/N)
• PRs: [rossoctl/<repo>#N](https://github.com/rossoctl/<repo>/pull/N)

Examples:

• [rossoctl/rossoctl#960](https://github.com/rossoctl/rossoctl/issues/960)
• [rossoctl/cortex#239](https://github.com/rossoctl/cortex/pull/239)

Phase 2: Deep Dive per Active Repo

For each repo classified as active in org-summary.json, analyze in order of activity level.

For rossoctl (local checkout available) — Full Depth

This is the main repo with local checkout. Apply the full github:last-week analysis:

1. Issue Analysis: For every open issue, search the local codebase for affected code/component. Check if a fix was merged. Classify severity (security, blocking, bug, feature, epic, stale).
2. PR Analysis: For every open PR, check CI status, review status, staleness, conflicts. Classify health (ready to merge, needs review, needs /run-e2e, CI failing, stale, conflicts).
3. CI Failure Timeline: Map failures on main to triggering commits. Identify recurring vs one-off failures. Correlate with merged PRs between last success and failure.
4. Root Cause Correlation: For each CI failure, identify candidate PRs by checking file paths touched (charts/, deployments/, .github/, rossoctl/backend/).

Use subagents for log analysis:

Agent(subagent_type='Explore'):
  "Read /tmp/rossoctl/github/data/rossoctl/merged-prs.json and summarize:
   - Count by author
   - Which areas changed most (charts, backend, tests, CI)
   Return a brief summary, not the raw data."

For Other Active Repos (API-only) — Moderate Depth

No local checkout. Use gathered JSON data only.

Use absolute URLs for all references (see Linking Convention above).

1. Issue Triage: Read open-issues.json. Group by labels. Flag stale issues (>30 days no update). Note issues linked to PRs.
2. PR Status: Read open-prs.json. Check CI/review status from JSON. Flag PRs waiting >7 days for review.
3. CI Health: Read ci-runs.json. Count pass/fail. Flag any failures on main.
4. Merged PR Summary: Read merged-prs.json. List with author and title.

Present each active repo as:

## <repo> (Deep Dive)

### Merged PRs (N)
| # | Title | Author | Merged |
| [rossoctl/\<repo\>#N](https://github.com/rossoctl/\<repo\>/pull/N) | ... | ... | ... |

### Open PRs (N)
| # | Title | Author | CI | Review | Health |
| [rossoctl/\<repo\>#N](https://github.com/rossoctl/\<repo\>/pull/N) | ... | ... | ... | ... | ... |

### Open Issues (N)
| # | Title | Labels | Age | Status |
| [rossoctl/\<repo\>#N](https://github.com/rossoctl/\<repo\>/issues/N) | ... | ... | ... | ... |

### CI Health
- Main branch: X/Y passed (Z%)
- Failures: [list if any]

Phase 3: Quiet Repos

For repos with no merged PRs and no new issues, present a single table:

## Quiet Repos

| Repo | Open PRs | Open Issues | Last Activity | Note |
|------|----------|-------------|---------------|------|
| repo-x | 0 | 2 | 45 days ago | dormant |
| repo-y | 1 | 0 | 12 days ago | low activity |

Check the oldest issue/PR update date to determine "last activity".

Phase 4: Cross-Repo Insights

Analyze patterns across all repos:

1. Shared Contributors: Which authors contributed to multiple repos this week?
2. Related PRs: Any PRs in different repos that reference each other or the same issue?
3. Org CI Health: Overall pass rate across all repos. Any repo dragging down the average?
4. Dependency Patterns: Any repo changes that might affect others (shared charts, common libs)?

Present as bullet list:

## Cross-Repo Highlights

- **@author** contributed to rossoctl (3 PRs) and repo-x (1 PR)
- Org CI pass rate: X% (rossoctl: Y%, repo-x: Z%)
- [any notable cross-repo patterns]

Phase 5: Generate Consolidated Report

Write the full report to /tmp/rossoctl/github/org-weekly-report.md:

# Rossoctl Org Weekly Report: [start-date] - [end-date]

## Org-Wide Summary
[table from Phase 1]

## Cross-Repo Highlights
[bullets from Phase 4]

## rossoctl (Deep Dive)
[full analysis from Phase 2]

## <other-active-repo> (Deep Dive)
[moderate analysis from Phase 2]

## Quiet Repos
[table from Phase 3]

## Action Items

| # | Action | Repo | Owner | Priority |
|---|--------|------|-------|----------|
| 1 | [highest priority] | rossoctl | @author | P0 |
| 2 | [next action] | repo-x | @author | P1 |
...

Action items are a single flat list across ALL repos, ordered by priority. No timescales.

Save the report:

# Write to /tmp/rossoctl/github/org-weekly-report.md

Phase 6: Ask User

After generating the report, ask:

Org-wide weekly report ready at /tmp/rossoctl/github/org-weekly-report.md. Want me to create a GitHub issue in $OWNER/$REPO with this report? Suggested title: "Org Weekly Report [start-date] - [end-date]"

You can also suggest a different repo or title.

Only create the issue after user confirms title, repo, and content.

Context Budget

Follow the CLAUDE.md context budget rules strictly:

• Redirect large output: The Phase 0 script output goes to a log file
• Use subagents for analysis: Never read full JSON files in the main context. Use Explore subagents to extract summaries.
• Small output OK inline: jq length, jq .totals, repo counts — these are fine inline.

Related Skills

• github:last-week - Deep single-repo report (used as the template for rossoctl deep dive)
• github:issues - Deep dive into individual issue triage
• github:prs - Deep dive into individual PR health
• ci:status - Detailed CI check analysis
• rca:ci - Investigate CI failures