allanbian1017 Daily Distiller
This skill automates reviewing today's intelligence reports and synthesizing them into a high-quality Knowledge Distillation document — capturing what's technically significant, how concepts connect, and what actions follow.
Workflow
Step 1: Discover Today's Reports
Find all .md files created today in the ./reports/ directory, excluding reports/distillations/. Use today's local date.
find reports -not -path "reports/distillations/*" -type f -name "*.md" -newermt "YYYY-MM-DD" ! -newermt "YYYY-MM-DD+1"
If no reports are found, notify the user and stop.
Step 2: Read All Reports
Read every report file in full. Note:
- Source / Author: newsletter name, Threads handle, etc.
- Key technical insights: Especially those about AI Agents, Context Engineering, Workflow Optimization, and System Architecture.
- Cross-cutting themes: Patterns that appear across multiple sources deserve to become a Pillar.
Step 3: Load the Output Template
Read assets/output_template.md — it contains the file output rules (path, filename format), document structure guide, and the full markdown template to follow.
Step 4: Synthesize
Check data/lang_preferences.md for the Preferred Report Language configuration, and produce the distillation document in that language (defaulting to English if not configured). Follow the template structure:
- Executive Summary — Overarching narrative. What is the "big picture theme" of today?
- Major Pillars (2–4 sections) — Group insights into coherent technical themes (not one per source). Name each section for the concept, not the source.
- The Synthesis — Explain how the pillars relate to each other as a system, not just a list.
- Strategic Takeaway — One actionable, philosophical call-to-action grounded in the day's data.
Quality bar: Avoid one-to-one report-to-section mapping. The goal is to reframe raw inputs into a unified technical narrative. The most valuable distillations surface connections that weren't obvious in the individual reports.
Step 5: Save Output
Follow the file output rules in assets/output_template.md:
reports/distillations/Knowledge_Distillation_YYYY_MM_DD.md
Use today's local date in the filename.
Principles (from docs/workflow/synthesis.md)
- System over Model: Prioritize insights that improve the system (context, workflows, rules), not just the model output.
- Explicit vs. Implicit: Actively identify "Explicit Knowledge" that can be codified into AI Skills vs. "Implicit Knowledge" that remains a human edge.
- Compounding Experience: High-quality outputs become templates. Treat this document as a long-term asset.