morning-brief

name	morning-brief
description	Formal protocol for executing the daily Morning Brief. Use this to actively mine specific YouTube channels and repos for gold nuggets applicable to open source.
type	encoded-preference

Morning Brief Protocol (Encoded Preference)

This skill formalizes an Encoded Preference for executing the "Morning Brief". CRITICAL: A Morning Brief is never a system status report. It is a highly deterministic intelligence-gathering operation against a specific lineup of external content creators to mine for "gold nuggets" (patterns, tools, architectural shortcuts) that can be applied to projects.

🎯 The Objective

Scan the latest videos, blog posts, and repository commits from the predefined Creator Lineup. Extract high-value, actionable insights that align with benchmark-agnostic routing and local-first architecture.

👥 Phase 1: Target Acquisition (The Lineup)

You must monitor the following specific sources:

Creator / Org	Key Works / Channels	Focus Area
openclaw	openclaw.ai, docs.openclaw.ai	MCP-first control planes, channel routing patterns
agent0ai	YouTube: @AgentZeroFW	Python agent frameworks, skill patterns
SEO Goldie	YouTube: @JulianGoldieSEO	Discovering new MCPs, free APIs, and affiliate setups
Nate B Jones	YouTube: @NateBJones	Structural security, architecture, daily monitors
Indy Dev Dan	YouTube: @IndyDevDan	Independent developer workflows, AI tool chains
AI Jason	YouTube: @AIJason	AI Agent workflows, LLM evaluations, product breakdowns
Fahd Mirza	YouTube: @FahdMirza	Enterprise AI tools, local LLM evaluation
Nick Ponte	YouTube: @NickPonte	SEO, web agency AI automations
Tina Huang	YouTube: @TinaHuang1	Productivity optimization, data workflows
Better Stack	YouTube: @BetterStack	Uptime monitors, synthetic testing patterns
Julia McCoy	YouTube: @JuliaMcCoy	Content engines, structural narrative AI
Matthew Berman	YouTube: @matthew_berman	AI news, open source model tests, Lindy OS
TheAIGRID	YouTube: @TheAIGRID	High-level architectural analysis, paper breakdowns
Two Minute Papers	YouTube: @TwoMinutePapers	State of the art computer graphics and AI research
Prompt Engineering	YouTube: @PromptEngineering	Structuring LLM contexts and constraints
Tool Drop	YouTube: @ToolDrop	Bleeding edge new tools
Matt Wolfe	YouTube: @mreflow	FutureTools API, SaaS platform developments
HuggingFace	YouTube: @HuggingFace	Open weights, Space definitions, Transformers
Ben AI / Nate Herk	YouTube: @BenAI92, @nateherk	Autonomous business processes, integration pipelines
Wes Roth	YouTube: @WesRoth	AI paradigm shifts, Auto-Research swarms, model evolution
Others	Hugo Wong, David Ondrej, theMITmonk, Nick Ai	General agentic developments and news

🛠️ Phase 2: Execution Steps

When the user asks for the "Morning Brief," execute exactly:

1. Acknowledge Intent: Confirm you are initiating the external mining protocol, NOT a system status report.
2. Scan Sources: Use search tools or browser subagents to check the YouTube channels and GitHub repos listed above for updates.
3. Check GitHub Repositories: Aggressively scan and check all relevant GitHub repositories associated with the creators and organizations listed in Phase 1 for actionable diffs/updates. This explicitly includes OpenClaw and its competitors:
   • openclaw/openclaw
   • agent0ai/agent-zero
   • anthropics/claude-code
   • xjtulyc/MedgeClaw
   • TinyAGI/tinyclaw
   • zeroclaw-labs/zeroclaw
   • BlockRunAI/ClawRouter
   • sipeed/picoclaw
   • HKUDS/ClawWork
   • RightNow-AI/openfang (Use tools like repomix to fetch README.md, CHANGELOG.md, RELEASES.md, and docs/**/*.md to identify structural changes).
4. Extract Nuggets: Identify specific technical patterns, new free API tiers, or architectural shortcuts mentioned.
5. Parse Descriptions & Comments: Use yt-dlp to download video descriptions and video comments. Parse them for hidden GitHub URLs, tool links, or additional context dropped by the creator or community.
6. Filter for Relevance: Discard generic hype. Keep only "gold nuggets" applicable to OpenDigz.
7. Format the Brief: Present findings concisely in morning-brief.md.

🚨 Trigger Tuning & Evlas

• Trigger: Do NOT activate on requests like "What is our status today?" or "Summarize git logs". ONLY activate on "Run the morning brief" or "Mine YouTube intel".
• Time/Latency Eval: This workflow requires heavy tool usage. Allow up to 2-3 minutes of execution time.
• Pass Rate: Must accurately extract transcripts without hallucinating content. If a creator hasn't posted, state there is no new intel. Do not invent insights.