Seventh step. Winner post = validated idea (resonance already proven) → reworking into other formats/angles has the highest EV. Writes recycle ideas into the Idea Pool, which then go through the regular pipeline 04→05→06. This closes the content cycle (compounding).
Input: $ARGUMENTS — client + mode. --backfill (default at start): scrape top-30/50
posts for the year. --from-tracking (monthly): fresh winners from Posts DB.
What we read
- Winners:
- backfill — Apify
apimaestro/linkedin-profile-posts(no-cookie), top by engagement (reactions + 2×comments) for the year; take top--top(default 30). - from-tracking — Posts DB (populates
content-track):Performance tier∈ {gem, strong},Recycle eligible= true. Priority — quadrant 🎯 "quiet lead-gen" (low reach, but produced dialogues): reworking a stronger format gives the most. 📣 "reach without leads" is not multiplied by format — we rewrite the angle for the ICP. Posts DB empty → say so directly and go to--backfill.
- backfill — Apify
- Strategy (Notion): pillars (map the repurposing), voice, offer ladder (spot lead-magnet candidates), fact-patrons.
- Recycling rules:
${CLAUDE_PLUGIN_ROOT}/reference/methodology.md§2.5a (8 angles), §2.6 (reuse), §4 (format conversion 45-60 days, republish ≥ 90 days).
Repurpose Matrix (output-visualization for each winner)
For each winner build heatmap angles × formats with a virality score in the cell (HTML widget via visualize, brand-node coral; legend + top-5 moves at the bottom).
- Rows (angles, from the post-generator / framework): Personal story · Lessons · Mistakes · How-I · Listicle · Case/social proof · Villain/contrarian.
- Columns (formats): Carousel · Lead-magnet · Infographic · Newsletter · Single image · Cross-post.
- Cell score = round(100 × format_mult × angle_weight × validated_factor):
format_mult(van der Blom 2026 ER): carousel/document 1.0 · lead-magnet 0.95 (giveaway-lever) · infographic 0.9 · newsletter 0.6 · single image 0.55 · cross-post 0.5.angle_weight— viral lever, calibrated to the author's data (content-track --mode=calibrate, Section 7 of the strategy) (with Viktor personal/vulnerability beats theory): pick weights for the DNA of the specific winner (for a personal-winner: Personal 1.0, Lessons/Mistakes 0.9, How-I 0.8, Listicle 0.65, Case 0.6, Villain 0.55).validated_factor— performance percentile of the original (top-1 ≈ 0.95).
- Pauses (cut off): original cell = republish (only ≥ 90 days); format conversion ≥ 45-60 days.
- Top-N cells → recycle cards. "Same format / adjacent topic" — a separate note line (republish ≥90d or a neighboring angle of the same topic).
Output
- Repurpose Matrix (heatmap) for each top winner — for review and selection.
- Recycle cards in 💡 Idea Pool (from selected/top cells):
Source = recycle·Source link= original URL ·Last used= original date (so the pauses work) ·Notes= "what won + what rework (angle×format, score)" · pillar/funnel/format of the new version · Score from the matrix. - Digest: top winners + matrices; what was filtered out due to a pause.
- Summary to the orchestrator (cards ready for content-weekly-plan).
Definition of Done
- Winners collected (backfill scrape or Posts DB winners); ranked by engagement.
- Each has 2-4 reworks from the matrix; pauses (45-60d conversion / ≥90d republish) observed.
- Recycle cards in Idea Pool with a link to the original and Last used = original date.
- Lead-magnet candidates (framework posts) marked separately (unlock giveaway).
- Digest shown; no duplicates against the existing pool created.