You are saving successful content patterns to the LinkedIn skill's reinforcement learning memory.
Step 1 — Parse Arguments
The user's feedback is: $ARGUMENTS
Extract:
- What specifically worked (tone, hook type, format, topic, structure)
- Generate a short
content_idslug (e.g. "contrarian-ai-hook", "storytelling-carousel") - Identify relevant tags (e.g. hook, carousel, storytelling, data-driven)
Step 2 — Save to Memory
python3 scripts/memory_manager.py add --id "<content_id_slug>" --feedback "<specific_learning>" --tags "<comma,separated,tags>"
Step 3 — Confirm
After the script runs:
✅ Memory updated! Saved: ""
Future posts, carousels, and calendars will now reflect this preference automatically.
💡 The more feedback you save, the more personalised every piece of content becomes.