agenda-intelligence-md 0.5.0

A drop-in markdown cognition layer for AI agents that need to analyze public agenda

Agenda Intelligence MD

A drop‑in cognition layer for AI agents that stops them from just summarizing news and starts producing decision‑ready analysis.

Bottom line: agents using this protocol move from “monitor developments” to “watch these 3 indicators; if X happens, decision Y becomes urgent.”

---

🚀 Quick start (OpenClaw)

1. Install skill (recommended):

   clawhub install agenda-intelligence
   

   Or copy skills/agenda-intelligence/ into your workspace.

2. Use in a prompt:

   [skill:agenda-intelligence]
   Brief: "EU AI Act amendment adds new high‑risk categories."
   

   The agent will load the protocol, run the CLI, and return a compact brief.

3. Or use the CLI directly (after pip install agenda-intelligence-md):

   agenda-intelligence validate-brief examples/agenda-brief.json
   agenda-intelligence source-plan technology-ai
   

---

🎯 What it does

Without Agenda‑Intelligence.md	With Agenda‑Intelligence.md
“Companies should monitor developments.”	“Watch for regulator guidance, first enforcement action, compliance deadline, and product redesigns. Treat as signal until those indicators appear.”

The protocol forces the agent to answer:

• Signal classification: noise → weak signal → signal → structural shift → trigger event
• What changed? (fact)
• Why it matters? (incentives, leverage)
• Who is affected?
• Main uncertainty?
• Scenarios & watch‑next indicators

Result: shorter output, higher decision value.

---

🧠 Key features

• Source Acquisition Layer – tells the agent which source types to check before making claims (sanctions, regulation, elections, conflict, etc.).
• AnalysisBank – a memory layer that stores reusable reasoning patterns from good/bad outputs.
• Regional lenses – Central Asia & Caspian, Middle East, EU.
• Sector lenses – sanctions, export controls.
• JSON schemas – validate briefs, evidence packs, memory cards.
• CLI & Python API – agenda-intelligence command, agent-manifest.json for discovery.

---

📦 Installation

# From PyPI (when published)
pip install agenda-intelligence-md

# Or editable install from source
git clone https://github.com/vassiliylakhonin/agenda-intelligence-md
cd agenda-intelligence-md
pip install -e .

---

🛠 CLI examples

agenda-intelligence --help
agenda-intelligence manifest
agenda-intelligence list-lenses
agenda-intelligence source-types
agenda-intelligence source-plan sanctions
agenda-intelligence validate-brief examples/agenda-brief.json
agenda-intelligence validate-evidence examples/source/evidence-pack.json
agenda-intelligence memory-search "sanctions routing"

---

📚 Documentation

• Protocol: Agenda-Intelligence.md – the core reasoning workflow.
• Quickstart: docs/quickstart.md
• Integrations: docs/integrations/ – Claude Code, OpenAI Codex, Cursor, MCP.
• Evaluation: docs/evaluation.md – how the scoring works.
• Roadmap: ROADMAP.md

---

🏗 Repository structure

agenda-intelligence-md/
├─ src/agenda_intelligence/   # Python package
├─ schemas/                   # JSON schemas
├─ examples/                  # sample briefs, evidence packs
├─ analysis-bank/             # memory cards (failures & successes)
├─ skills/agenda-intelligence/ # OpenClaw skill wrapper
├─ docs/                      # guides & integration notes
└─ tests/                     # pytest suite

---

🤝 Contributing

Pull requests are welcome! Please:

1. Open an issue to discuss changes.
2. Create a feature branch (feat/..., fix/...).
3. Run pytest and ensure all tests pass.
4. Update docs if behavior changes.

See docs/quickstart.md for development setup.

---

📄 License

MIT – free for any use.

---

🔗 Related projects

• global-think-tank-analyst – deeper policy‑risk memos for OpenClaw.
• ReasoningBank – the inspiration for AnalysisBank.

---

Why this exists: Most agent‑written news analysis is a polished recap that doesn’t change any decision. This project gives agents a stricter workflow so their output actually helps someone decide, hedge, or act.