v0.1.0 — Agenda-Intelligence.md

First stable release of Agenda-Intelligence.md as a drop-in cognition layer for agents that analyze public agenda.

What is included

• Top-level Agenda-Intelligence.md file designed to sit next to AGENTS.md.
• ADOPTION.md with setup patterns for repo instructions, AGENTS.md, retrieval/RAG, and OpenClaw-style skills.
• Base agenda-analysis protocol:
  • signal classification;
  • fact / assessment / assumption / unknown / scenario / indicator separation;
  • evidence discipline;
  • output patterns.
• Before/after demos showing the difference between generic summary and decision-useful agenda analysis.
• Evaluation rubric for scoring agent outputs.
• Regional lens packs:
  • Central Asia + Caspian;
  • Middle East;
  • European Union.
• Sector lens pack:
  • Sanctions.
• Optional OpenClaw-compatible skill wrapper.
• CI validation for required files.

Recommended use

Copy Agenda-Intelligence.md into an agent repo and add a conditional hook to AGENTS.md or equivalent instructions:

When analyzing public agenda, news, policy, regulation, sanctions, geopolitics, trade, elections, conflicts, markets, or strategic risk, follow `Agenda-Intelligence.md`.

Do not summarize by default. Classify the signal, identify what changed, separate fact from assessment, name uncertainty, and end with watch-next indicators.

Why it exists

Generic agents often produce:

recap → generic implications → “monitor developments”

This release pushes agents toward:

signal classification → what changed → affected actors → uncertainty → scenarios → watch-next indicators