Content Mine

Content Mining

Overview

Scan recent activity across repos to surface content ideas. Create content queue issues for anything worth posting about.

Sources

Two parallel signal sources feed the content queue:

1. Git Activity (manual / scheduled)

Run this skill to scan recent dev work. Default time range: last 7 days. User can specify: "mine the last 3 days", "mine since Monday".

Git commits (key repos):

for repo in waaseyaa/framework waaseyaa/giiken jonesrussell/blog jonesrussell/jonesrussell; do
  echo "=== $repo ==="
  gh api repos/$repo/commits --jq '.[].commit | "\(.author.date) \(.message)"' --paginate 2>/dev/null | head -20
done

Closed issues and milestones:

for repo in waaseyaa/framework waaseyaa/giiken jonesrussell/jonesrussell; do
  echo "=== $repo ==="
  gh issue list --repo $repo --state closed --json number,title,closedAt --limit 10
done

New design specs:

find ~/dev/blog/docs/superpowers/specs/ -name "*.md" -newer /tmp/last-mine-marker 2>/dev/null

Blog posts published:

find ~/dev/blog/content/posts/ -name "index.md" -newer /tmp/last-mine-marker 2>/dev/null

2. North Cloud Classified Content (event-driven / automatic)

A separate signal handler service subscribes to North Cloud's Redis pub/sub and creates content-queue issues automatically. These issues arrive with source:north-cloud label.

• Channel: coforge:core (developer/startup content classified as core relevance)
• Filters: quality_score ≥ 65, content_type article or blog_post, deduped by URL
• Output: stage:mined, source:north-cloud, type:text-post
• Brand filter: only software/AI/builder ecosystem content — no crime, entertainment, mining, indigenous

This source runs independently. When you run /content-mine, North Cloud items already in the queue appear alongside git-activity items. You don't need to trigger them manually.

Process

3. Filter noise. Skip:

   • Merge commits, dependency bumps, CI fixes
   • Commits with messages under 10 characters
   • Typo fixes, formatting-only changes
   • Anything already in the content queue (check existing issues)

4. Assess each candidate. For each remaining item, ask:

   • Does this demonstrate something interesting? (feature, technique, milestone)
   • Would this resonate with the audience? (AI, Waaseyaa, open source, PHP, Go, software fundamentals)
   • Is there enough substance for a post?

5. Create queue issues. For each surfaceable item:

   gh issue create --repo jonesrussell/jonesrussell \
     --title "[content] <descriptive title>" \
     --label "content-queue,stage:mined,type:<suggested-type>" \
     --body "<filled content queue template>"
   

6. Update marker.

   touch /tmp/last-mine-marker
   

7. Report. Show what was mined:

   • N items surfaced (with issue numbers)
   • Key themes spotted
   • Suggest running /curate to review

Noise Filtering Heuristics

Over-Surface

When in doubt, create the issue. It's easier to skip during curation than to miss something worth posting about.