An MCP/HTTP/A2A runtime that takes evidence packs for strategic-intelligence and due-diligence workflows and returns structured trust-routing decisions, evidence gaps, audit data, and human-review escalation. It ships several deployed vertical workers, including Kazakhstan deal-risk and market-entry readiness gates plus CIS secondary-sanctions exposure triage.
From the registry: Auditable strategic-risk memos for AI agents: sanctions, regulatory, geopolitics, trade.
$ pip install agenda-intelligence-mdhttps://pypi.org/project/agenda-intelligence-md/$ pip install "agenda-intelligence-md[llm]"$ export ANTHROPIC_API_KEY=<your-api-key>https://console.anthropic.com/$ export OPENSANCTIONS_API_KEY=<your-api-key>https://www.opensanctions.org/api/Please install the `agenda-intelligence-md` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install the Python package** — Install the MCP server package from PyPI. Run: `pip install agenda-intelligence-md` (https://pypi.org/project/agenda-intelligence-md/)
Optional prerequisites:
- Optional LLM support dependencies — Only needed if you want the `analyze` tool to call Anthropic directly instead of letting the host model complete from the returned system prompt. Run: `pip install "agenda-intelligence-md[llm]"`
- Anthropic API key — Required only when using the optional LLM-backed `analyze` mode with Anthropic. Run: `export ANTHROPIC_API_KEY=<your-api-key>` (https://console.anthropic.com/)
- Optional Watchman service — Optional upstream for per-profile live retrieval in the CIS secondary-sanctions worker. Self-host `moov-io/watchman` and set `WATCHMAN_URL` to enable it. (https://github.com/moov-io/watchman)
- Optional OpenSanctions API key — Optional fallback upstream for per-profile live retrieval in the CIS secondary-sanctions worker. If both Watchman and OpenSanctions are configured, Watchman takes precedence. Run: `export OPENSANCTIONS_API_KEY=<your-api-key>` (https://www.opensanctions.org/api/)
Canonical MCP server config (stdio transport):
- command: `agenda-intelligence`
- args: []
- optional environment variables:
- `ANTHROPIC_API_KEY`: Anthropic API key for optional LLM-backed `analyze` execution when using the `[llm]` extra. (example: `<your-api-key>`)
- `WATCHMAN_URL`: URL of a self-hosted Watchman instance to enable optional live retrieval for the `cis_secondary_sanctions_exposure` worker. (example: `<your-watchman-url>`)
- `OPENSANCTIONS_API_KEY`: OpenSanctions API key to enable optional live retrieval fallback for the `cis_secondary_sanctions_exposure` worker when Watchman is not configured. (example: `<your-api-key>`)
Note: The README documents a stdio MCP server and says the full product behavior remains in the installable stdio MCP server. No client-wrapper JSON is included here; this is the canonical command/args/env shape only. The package also ships separate HTTP and A2A surfaces, but the MCP install path is the stdio server via the `agenda-intelligence` command.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.ANTHROPIC_API_KEYNeeded only if using the optional LLM extra so the server can call the Anthropic API itself.WATCHMAN_URLEnables per-profile live retrieval for the CIS secondary-sanctions worker using a self-hosted Watchman upstream.OPENSANCTIONS_API_KEYEnables per-profile live retrieval for the CIS secondary-sanctions worker using the OpenSanctions API fallback.Meta Ads MCP (Facebook + Instagram) - analyze performance, manage budgets, pause campaigns.