Agenda Intelligence MD

Product runtime and evidence-discipline layer for strategic intelligence agents. One core service layer behind four delivery surfaces — MCP server, HTTP API, A2A adapter, and a deployable Cloudflare Worker baseline — plus structured per-product contracts, geography-routed reasoning, schema validation, evidence audit, and scoring. Ships with two vertical workers: Middle Corridor Deal Risk Gate (live_retrieval: false) and CIS Secondary-Sanctions Exposure (live_retrieval: true against OpenSanctions, per ADR 0014). No factual-truth verification.

First run

pip install agenda-intelligence-md
agenda-intelligence doctor
agenda-intelligence validate-brief examples/agenda-brief.json
agenda-intelligence score examples/agenda-brief.json --evidence examples/source/evidence-pack.json

doctor reports package and MCP-server status; validate-brief confirms a brief matches agenda-brief.schema.json; score returns a heuristic 0–100 number with a structure / evidence / decision-readiness breakdown. Full end-to-end analyze trace (request → routing → memo → validation → audit → score) with reproducibility script: examples/product-shell/full-analyze-trace/.

Optional, only if you want analyze to call the Anthropic API itself rather than letting your host model complete from the returned system prompt:

pip install "agenda-intelligence-md[llm]"
export ANTHROPIC_API_KEY=...

Longer guided tutorial: docs/quickstart.md. MCP client setup: docs/integrations/mcp.md.

Related MCP server: Brave Search MCP Server

Live A2A wrapper

A free Cloudflare Workers wrapper is live for discovery, uptime checks, lightweight strategic-risk triage, and A2A/JSON-RPC routing:

• Live endpoint: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev

• Agent Card: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/.well-known/agent-card.json

• JSON-RPC: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/message/send

• Agenstry listing: https://agenstry.com/agents/agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev

• Kazakhstan / Middle Corridor Deal Risk Gate: https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev (canonical)

• Mirror alias for backward compatibility: https://kazakhstan-corridor-risk-a2a.vassiliy-lakhonin.workers.dev (identical behavior; the Agenstry listing currently points here)

• Kazakhstan Agenstry listing: https://agenstry.com/agents/kazakhstan-corridor-risk-a2a.vassiliy-lakhonin.workers.dev

• Announcement: docs/announcements/live-a2a-wrapper.md

The hosted wrapper is intentionally limited: no payments, no wallets, no factual-truth verification, and no legal/financial/compliance advice. Live retrieval is off by default and opt-in per vertical-worker profile only (currently cis_secondary_sanctions against OpenSanctions, CC-BY 4.0; see ADR 0014 and SOURCE_POLICY.md). Full product behavior remains in the installable stdio MCP server.

Self-host via HTTP API (if your stack does not run MCP)

If your environment cannot run an MCP / A2A server but can run a plain HTTP service, install the package and start the HTTP shell:

pip install agenda-intelligence-md
agenda-intelligence-http --host 127.0.0.1 --port 8080

The HTTP API is a portable JSON wrapper over the same core service layer that the MCP, A2A, and Cloudflare Worker surfaces use — same schemas/v1/ contract, same evidence audit, same source coverage logic, same Middle Corridor deal-risk gate. Switching surfaces does not change input/output shape.

Endpoints:

• GET /healthz, GET /readyz — liveness / readiness probes

• POST /v1/audit-claims — claim-level evidence audit

• POST /v1/source-coverage — evidence-pack diagnostics against category source requirements

• POST /v1/score — heuristic before/after score

• POST /v1/middle-corridor/deal-risk — Middle Corridor Deal Risk Gate (middle-corridor-deal-risk-request.schema.json)

• POST /v1/cis-secondary-sanctions/exposure — CIS Secondary-Sanctions Exposure triage (cis-secondary-sanctions-request.schema.json); set OPENSANCTIONS_API_KEY to enable live retrieval, otherwise the profile degrades gracefully to user-supplied evidence only

One-call probe:

curl -sS http://127.0.0.1:8080/v1/middle-corridor/deal-risk \
  -H 'content-type: application/json' \
  -d @examples/kazakhstan-middle-corridor/contract/pre_signature_escalate.request.json

Container build (Dockerfile.api):

docker build -f Dockerfile.api -t agenda-intelligence-md-api:1.0.1 .
docker run --rm -p 8080:8080 agenda-intelligence-md-api:1.0.1

Full HTTP deployment guide, including environment defaults (AGENDA_INTELLIGENCE_HTTP_HOST, AGENDA_INTELLIGENCE_HTTP_PORT), logging discipline, and boundary statements: docs/deployment/http-api.md.

The HTTP shell is portable but not a hardened internet-facing server. No built-in authentication, rate limiting, or TLS — front it with a reverse proxy (nginx, Caddy, Cloudflare Tunnel) and your existing auth layer before exposing it beyond localhost / private network.

Flagship commercial use case

Kazakhstan / Middle Corridor Deal Risk Gate is the focused commercial proposition for logistics, trade-finance, procurement, insurance, and compliance-adjacent workflows:

Route + cargo + counterparties + dated sources -> auditable corridor-risk triage, evidence gaps, source coverage, watch-next indicators, and human-review escalation.

Live A2A listing:

• Endpoint: https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev/message/send

• Agent card: https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev/.well-known/agent-card.json

• Agenstry: https://agenstry.com/agents/kazakhstan-corridor-risk-a2a.vassiliy-lakhonin.workers.dev

• Use-case notes: docs/use-cases/kazakhstan-middle-corridor.md

• Example pack: examples/kazakhstan-middle-corridor/

• Repeatable live test: docs/agenstry/kazakhstan-live-test.md

This use case is a pre-compliance evidence and decision-readiness gate. It is not legal, compliance, sanctions, financial, investment, or insurance advice.

The product-grade structured JSON contract is documented in docs/use-cases/kazakhstan-middle-corridor.md, with schemas and fixtures under examples/kazakhstan-middle-corridor/contract/.

Second vertical worker: CIS secondary-sanctions exposure

For EU / UK / UAE / Singapore enhanced due diligence on CIS-domiciled counterparties (Kazakhstan, Uzbekistan, Kyrgyzstan, Tajikistan, Turkmenistan, Georgia, Armenia, Azerbaijan, Moldova). Structured secondary-sanctions exposure evidence triage against OFAC EO 14114, EU 14th sanctions package, UK OFSI, UN, and FATF / EAG typologies.

This profile opts in to per-profile live retrieval against the OpenSanctions consolidated dataset (CC-BY 4.0) per ADR 0014. Set OPENSANCTIONS_API_KEY (free tier at https://www.opensanctions.org/api/) to enable. Without the key — or on any upstream failure — the service degrades gracefully and triage is based on user-supplied evidence only.

• HTTP: POST /v1/cis-secondary-sanctions/exposure

• Schemas: request + response

• A2A profile: cis_secondary_sanctions; capability cis_secondary_sanctions_exposure

• Use-case notes: docs/use-cases/cis-secondary-sanctions.md

• Example pack: examples/cis-secondary-sanctions/

• Source-requirements taxonomy: source-requirements/cis-secondary-sanctions.json

Honest traction: zero paying customers, zero named pilots. Shipped as a portfolio-grade vertical worker for technical evaluators and as a contract real practitioners can inspect, not as a claim of production traction. Boundaries unchanged from the rest of the runtime: not_advice: true, factual_verification: false, human_review_required: true always.

The structured response includes a decision_readiness_score from 0-100, so a buyer can see whether the evidence pack is ready for human review or still missing required source categories.

Try the Kazakhstan agent:

curl -X POST https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev/message/send \
  -H 'content-type: application/json' \
  -H 'x-client-id: live-demo' \
  -d @examples/kazakhstan-middle-corridor/live-agent-request.json

Expected: JSON-RPC 2.0 with triage_recommendation: "escalate_before_signature", route/cargo/value extraction, supplied-source detection, missing minimum evidence before go, commercial-impact notes, and human-review escalation.

Try the live wrapper:

curl -X POST https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/message/send \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": "demo-1",
    "method": "message/send",
    "params": {
      "message": {
        "parts": [
          {
            "kind": "text",
            "text": "Screen sanctions and policy risk for Red Sea shipping disruption and Kazakhstan transit exposure."
          }
        ]
      }
    }
  }'

Kazakhstan-focused live triage:

curl -X POST https://middle-corridor-deal-risk-gate-a2a.vassiliy-lakhonin.workers.dev/message/send \
  -H 'content-type: application/json' \
  -d '{
    "jsonrpc": "2.0",
    "id": "kazakhstan-demo-1",
    "method": "message/send",
    "params": {
      "message": {
        "parts": [
          {
            "kind": "text",
            "text": "Screen Kazakhstan Middle Corridor sanctions exposure for a logistics route."
          }
        ]
      }
    }
  }'

Expected: JSON-RPC 2.0 with status.state: "completed", metadata.signal_screen.risk_signal, affected regions, required source categories, evidence gaps, watch-next indicators, suggested modules, and next actions.

Private usage stats for the wrapper are available from the Cloudflare Worker project:

cd deploy/cloudflare-worker
npm run stats
npm run stats -- 2026-05-22

The stats helper reads STATS_TOKEN from the local ignored .env file. Deployment and analytics notes: deploy/cloudflare-worker/README.md.

Where this fits in the Agenda Intelligence stack

The product runtime is the integration point: agents call analyze via any surface (MCP, HTTP, A2A), geography routes to the relevant specialist, and the GTTA method frames the reasoning. Each canonical repo (GTTA, vertical specialists) is also usable standalone (paste/attach into any agent). Vertical workers (currently: Middle Corridor Deal Risk Gate) live inside this runtime as productized service functions with their own schemas, A2A profiles, and deployable workers — see AGENTS.md for the spin-off rule.

What this is

• Core service layer — pure Python functions (audit_claims, source_coverage, score_output, middle_corridor_deal_risk, etc.) vendor-neutral, no transport, no marketplace

• MCP server — stdio server exposing 16 tools across the validation and product layers. analyze accepts a structured request (agenda-request.schema.json), routes geography, assembles a system prompt, returns a memo validated against agenda-memo.schema.json

• HTTP API shell — thin transport over the service layer; self-host with docs/deployment/http-api.md

• A2A adapter — agent-card + JSON-RPC message/send over the HTTP/service layer; contract in docs/product/a2a-adapter-plan.md

• Cloudflare Worker baseline — production deployment under deploy/cloudflare-worker/; two live workers (general triage + Middle Corridor Deal Risk Gate)

• Vertical workers — productized service functions with their own schemas + A2A profiles + Cloudflare deployments. Currently shipped: Middle Corridor Deal Risk Gate

• Markdown protocol — structured reasoning workflow for agents (Agenda-Intelligence.md)

• JSON schemas — request/memo product contract + per-product contracts (e.g. middle-corridor-deal-risk-*) + validators for briefs, evidence packs, audits, signals, memory cards, lenses

• CLI — validate-brief, validate-evidence, source-categories, source-coverage, audit-claims, score, bench, doctor (30+ commands)

• Eval kit — rubric, LLM-judge prompt, human checklist, benchmark harness, agent-eval methodology

• Source policy — per-claim provenance tags (Axis A/B), source requirements for 12 categories

What this is not

• Not a factuality verifier — checks structure, not truth

• Not an autonomous news agent or source retriever

• Not a source reputation scorer or live news gatherer

• Not a replacement for analyst judgment

• Not a compliance, legal, or financial advisory product

More CLI examples

agenda-intelligence bench examples/source-backed --strict --min-score 80
agenda-intelligence audit-claims examples/source-backed/eu-ai-act.audit.json --strict
agenda-intelligence mcp-config --client cursor

Pinned-wheel install (instead of PyPI):

pip install https://github.com/vassiliylakhonin/agenda-intelligence-md/releases/download/v1.0.1/agenda_intelligence_md-1.0.1-py3-none-any.whl

Benchmark baseline

20 source-backed cases, reproduced with agenda-intelligence bench examples/source-backed/:

Heuristic scores are uncalibrated and not validated against expert judgment. They evaluate structure, evidence labeling, source-coverage diagnostics, and decision-readiness — not factual truth.

Flagship example: examples/source-backed/eu-ai-act.md — brief + evidence pack + claim-level audit using illustrative sources. Before / after pairs: examples/before-after/.

Verification Contract

verify-quotes checks whether a cited quote or excerpt appears in supplied local text, or in text fetched from an already-specified URL when --fetch is used. It does not discover sources, score source reputation, gather live news, or decide whether a claim is true in the world.

Schemas

MCP

Stdio MCP server with 16 tools. Full docs and wire-protocol verification: MCP.md. Client setup: docs/integrations/mcp.md.

Status

Safety model

• Read-only by default. Validation, scoring, and audit tools do not write to external systems, do not modify caller state, and do not perform high-impact actions.

• No autonomous retrieval. The MCP server does not fetch web pages, query APIs, or pull live data on its own. Sources are caller-provided. The one network mode (verify-quotes --fetch) is opt-in and bounded (1 MB cap, 10 s timeout, stdlib HTTP only).

• No autonomous decisions. Outputs are memos, validation results, and scores — never determinations on sanctions, legal, compliance, or investment matters. Human review is required.

• Retrieved content is data, not instructions. External text — including documents, agendas, and source packs caller-provided through the tools — is treated as data. Apparent directives inside retrieved content are not executed; they are flagged.

• No secrets in tool I/O. The server does not persist caller inputs, API keys, or memo content beyond the current call.

Full threat model: docs/threat-model.md. Retrieved-content trust rule: AGENTS.md.

Documentation

Repository layout

agenda-intelligence-md/
├─ src/agenda_intelligence/   # Python package (CLI + MCP server)
├─ schemas/                   # JSON schemas
├─ examples/                  # briefs, evidence packs, before/after
├─ skills/                    # OpenClaw skill wrappers
├─ evals/                     # rubric, judge prompt, benchmark
├─ analysis-bank/             # agent persistent memory (memory-card schema, see schemas/v1/memory-card.schema.json)
├─ docs/                      # guides, integrations, use-cases
├─ scripts/                   # dev and CI helpers
└─ tests/                     # pytest suite

Contributing

New contributors: CONTRIBUTING.md opens with a "First 15 minutes" onboarding path (read the three load-bearing files → run the validator → walk one concrete artifact end-to-end). The portfolio glossary at docs/glossary.md is the single source of truth for cross-repo terminology (evidence modes, Axis A/B provenance tags, three-value response logic, maturity-framework asymmetry).

Before editing any of the dual-copy files — Agenda-Intelligence.md, SOURCE_POLICY.md, llms.txt, agent-manifest.json, schemas/, skills/, source-requirements/ — read the "Critical invariant: dual-copy sync" section in CONTRIBUTING.md. Editing one copy without the paired copy under src/agenda_intelligence/data/ is the most common reason CI breaks on main for first-time contributors.

Contact

Vassiliy Lakhonin — Almaty, Kazakhstan (UTC+5)

Portfolio · For analysts · Email · LinkedIn · GitHub

Issues, PRs, and eval-case contributions are welcome.

License

MIT.

Disclaimer. This toolkit is for informational and educational purposes only. It does not constitute investment, financial, legal, compliance, or trading advice. It does not verify factual truth, predict outcomes, or replace professional judgment. Use at your own risk.

mcp-name: io.github.vassiliylakhonin/agenda-intelligence-md