Agenda Intelligence MD

Agenda Intelligence MD is an evidence-readiness and trust-routing runtime for high-stakes AI-assisted decisions.

The sharpest current wedge is AI vendor evidence-readiness for regulated procurement: given an RFP, vendor documentation, model card, source pack, risk memo, or weekly project note, it produces a human-review packet:

• what is supported
• what is weak or unsafe to repeat
• what evidence is missing
• which owner action is next
• whether the file is ready for human review or must be escalated

Use it when an AI governance, procurement, third-party-risk, or analyst owner has to prepare a file for a near-term review and needs to know what cannot safely go to committee yet. The normal workaround is a manual checklist, spreadsheet, GRC ticket, consultant memo, or reviewer comments scattered across docs. Agenda Intelligence MD gives agents and local workflows a repeatable evidence-readiness contract around that work.

It does not verify factual truth, approve vendors, clear sanctions risk, provide legal/compliance/financial advice, or make autonomous decisions. It routes evidence gaps to humans. Current commercial status: build-to-learn / portfolio-proof; no production traction is claimed. Success means buyer behavior such as a redacted file, second review request, workflow correction, budget-owner intro, pilot, or payment -- not compliments.

Technically, one core service layer sits behind four delivery surfaces — MCP server, HTTP API, A2A adapter, and a deployable Cloudflare Worker baseline — with structured per-product contracts, geography-routed reasoning, schema validation, evidence audit, and scoring. Five deployed vertical profiles are available as portfolio/demo surfaces; they are proofs of contract shape, not proof of product-market fit. Evaluate them with live curl calls: docs/agenstry/demo-pack.md.

Why a repo full of markdown?

The skills under skills/ and the canon docs are source, not documentation — a SKILL.md is instructions an agent loads and executes, the way a runtime loads a module. The Python in src/ is the runtime that composes them: global-think-tank-analyst as the reasoning method, central-asia-caspian and gulf-middle-east as regional layers (module roles in llms.txt / agent-manifest.json). Both are version-controlled for the same reason — they're what runs.

Install in Claude Code

/plugin marketplace add vassiliylakhonin/agenda-intelligence-md
/plugin install agenda-intelligence@agenda-intelligence

This installs the MCP server (runs uvx --from agenda-intelligence-md agenda-intelligence-mcp; requires uv) plus the agenda-intelligence and source-ingest skills. The same marketplace also lists the method and regional skill plugins: global-think-tank-analyst, central-asia-caspian, gulf-middle-east.

For any other MCP client, add the server directly:

{
  "mcpServers": {
    "agenda-intelligence": {
      "command": "uvx",
      "args": ["--from", "agenda-intelligence-md", "agenda-intelligence-mcp"]
    }
  }
}

First run

If you are evaluating from Glama or another MCP directory, start with the local stdio server and one source-backed packet. The quickest path is validation/audit first, not a live vertical worker.

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
agenda-intelligence weekly-delta examples/strategic-infrastructure-bankability/status.synthetic.md

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; check-memo-quality and memo-quality-bench catch schema-valid but unsafe memo output; weekly-delta turns a redacted weekly/status note into unsafe-to-repeat claims, source-plan gaps, owner actions, and a decision-readiness route; weekly-delta-bench regression-tests that confidential workflow output. 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.

Live A2A wrapper

A free Cloudflare Workers wrapper is live for discovery, uptime checks, lightweight strategic-risk triage, and A2A/JSON-RPC routing. Six workers are deployed (general triage + the five vertical workers); per-worker endpoints live in each worker section below and in the Status table.

• Interactive browser demos: Middle Corridor Deal Risk Gate · CIS Secondary-Sanctions Exposure
• General wrapper: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev · agent card · JSON-RPC · Agenstry
• Agentic Resource Discovery catalog: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/.well-known/ai-catalog.json
• Confidential project-room profile contract: https://agenda-intelligence-a2a.vassiliy-lakhonin.workers.dev/profiles/confidential-project-room · synthetic redacted JSON
• Worked curl calls for selected live workers: docs/agenstry/demo-pack.md · repeatable Kazakhstan test: docs/agenstry/kazakhstan-live-test.md

Try a live portfolio demo:

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, the minimum evidence still missing before go, and human-review escalation.

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; it is currently active only for cis_secondary_sanctions, via the $0 Snapshot upstream (see ADR 0014 / ADR 0020 and SOURCE_POLICY.md). Private usage stats: deploy/cloudflare-worker/README.md. Core validation and product-shell behavior is available in the installable stdio MCP server; some deployed vertical profiles are HTTP/A2A-only.

Confidential project-room workflow

For confidential strategic-infrastructure and AI-compute project files, use the same evidence-readiness discipline without naming parties: docs/trust/confidential-project-workflow.md, the public profile contract at profiles/confidential-project-room/index.md, the schema at schemas/v1/confidential-project-room-profile.schema.json, docs/templates/strategic-infrastructure-evidence-readiness-profile.md, docs/templates/weekly-status-decision-readiness-delta.md, and the synthetic alias-only example pack at examples/strategic-infrastructure-bankability/. This is also build-to-learn: no new worker, no case-study claims, and no client names.

The deterministic local shell is available now:

agenda-intelligence source-plan ai-infrastructure-bankability
agenda-intelligence weekly-delta examples/strategic-infrastructure-bankability/status.synthetic.md

It converts status activity into a reviewer scaffold: status-to-evidence rows, new claim candidates, unsafe-to-repeat claims, owner actions, missing source-plan categories, and a route such as escalate_before_committee. It is keyword-based and deterministic, not full semantic analysis.

Write a Markdown packet:

agenda-intelligence weekly-delta examples/strategic-infrastructure-bankability/status.synthetic.md \
  --project-alias ProjectCo \
  --decision-moment "committee review" \
  --out readiness-delta.md

Live portfolio demo: Middle Corridor

Kazakhstan / Middle Corridor Deal Risk Gate is a live portfolio/demo profile 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.

The structured response also presence-flags sanctions-relevant / high-risk and re-export / circumvention-watch counterparty jurisdictions, decomposes risk into a domestic-legal vs foreign-sanctions exposure view, and surfaces a vessel deceptive-shipping-practice checklist for the maritime leg. All of this is presence-flagging and evidence triage routed to human review — not a sanctions determination. This profile is not assumed to be the commercial flagship; Kazakhstan/local-forwarder demand must be revalidated before further buyer-facing expansion.

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/middle-corridor-deal-risk-gate-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/.

Kazakhstan market-entry readiness contract

Kazakhstan Market Entry Readiness Gate is a structured contract pack for distribution, import, service, showroom, EPC, renewable-energy, infrastructure, technology-transfer, and partner-entry files:

Company + project + Kazakhstan objective + counterparties + supplied sources -> gate decision, readiness label, evidence gaps, claim audit, owner actions, watch-next indicators, and human-review routing.

It is a live vertical worker: a kazakhstan_market_entry_readiness service function, an HTTP route, an A2A profile, and a deployed Cloudflare Worker at https://kazakhstan-market-entry-readiness-a2a.vassiliy-lakhonin.workers.dev. It is HTTP/A2A-only in the current release, not exposed as a local stdio MCP tool. It is not legal, compliance, customs, tax, financial, investment, insurance, sanctions, or launch-authorization advice.

• Use-case notes: docs/use-cases/kazakhstan-market-entry-readiness.md
• Request schema: schemas/v1/market-entry-readiness-request.schema.json
• Response schema: schemas/v1/market-entry-readiness-response.schema.json
• Source taxonomy: source-requirements/kazakhstan-market-entry-readiness.json
• Example pack: examples/kazakhstan-market-entry-readiness/contract/
• HTTP: POST /v1/market-entry/readiness
• A2A profile: kazakhstan_market_entry_readiness
• Live endpoint: https://kazakhstan-market-entry-readiness-a2a.vassiliy-lakhonin.workers.dev · agent card

Second vertical worker: CIS secondary-sanctions exposure

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

This profile runs per-profile live retrieval with three upstream options, tried in order (free before paid), per ADR 0014 / ADR 0020:

1. Snapshot (active, $0, no host) — the worker fetches a compact public-list name index (OFAC SDN + consolidated, EU, UK FCDO) published by the portfolio site and matches the counterparty name in-worker (exact + token overlap). Activated via SNAPSHOT_INDEX_URL; deployed and live since 2026-06-26.
2. Watchman (free self-host) — moov-io/watchman Apache-2.0 self-host on an always-on container. Set WATCHMAN_URL to use instead.
3. OpenSanctions (paid) — hosted API at €0.10/call. Set OPENSANCTIONS_API_KEY (30-day business-email trial at https://www.opensanctions.org/api/, then per-call billing).

The first active upstream wins (Snapshot, unless SNAPSHOT_DISABLED=1). When none is configured or the active one fails, the service degrades gracefully and triage is based on user-supplied evidence only — live_retrieval_status: degraded / disabled in the response. A match is a possible string match only, not identity verification or a sanctions determination; human review is required. No paying customers — illustrative usage only.

• Live endpoint: https://cis-secondary-sanctions-a2a.vassiliy-lakhonin.workers.dev · Agenstry
• 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.

Third vertical worker: Agentic Interaction Trust Gate

For trust-and-safety, fraud-risk, product-security, and platform teams reviewing agent-mediated actions across checkout, account, API, MCP tool, and A2A endpoint surfaces.

This worker does not decide whether an actor is a bot. It asks whether the supplied evidence is sufficient to route a specific automated or agentic action: allow_low_risk, require_step_up, escalate_to_human_review, block_until_verified, not_decision_ready, or insufficient_information.

• Live endpoint: https://agentic-interaction-trust-a2a.vassiliy-lakhonin.workers.dev · Agenstry
• HTTP: POST /v1/agentic-interaction/trust
• A2A capability: agentic_interaction_trust
• Schemas: request + response
• Use-case notes: docs/use-cases/agentic-interaction-trust.md
• Example pack: examples/agentic-interaction-trust/
• Source-requirements taxonomy: source-requirements/agentic-interaction-trust.json

Boundaries: no cybersecurity monitoring, fraud adjudication, identity verification, transaction authorization, legal advice, compliance advice, or financial advice. The worker returns evidence gaps, readiness scoring, watch-next indicators, and human_review_required: true.

Fourth vertical worker: Gulf Maritime Exposure Gate

For trade-finance, marine-insurance, P&I, chartering, and compliance teams reviewing a vessel or voyage transiting the Strait of Hormuz, Persian/Arabian Gulf, Gulf of Oman, Bab-el-Mandeb, or Red Sea.

Structured triage of maritime sanctions and chokepoint-disruption exposure — Iran-oil, Russia price-cap, dark-fleet, STS transfer, flag-hopping, P&I gap, AIS manipulation, ownership/control, dual-use cargo — into an evidence-sufficiency routing decision: insufficient_information, escalate_before_fixture, escalate_before_voyage, not_decision_ready, or ready_for_human_review.

• HTTP: POST /v1/gulf-maritime/exposure
• A2A capability: gulf_maritime_exposure
• Live endpoint: https://gulf-maritime-exposure-a2a.vassiliy-lakhonin.workers.dev · Agenstry
• Schemas: request + response
• Use-case notes: docs/use-cases/gulf-maritime-exposure.md
• Example pack: examples/gulf-maritime-exposure/
• Source-requirements taxonomy: source-requirements/gulf-maritime-exposure.json

Boundaries: no live retrieval, does not resolve vessel ownership or verify identity, no legal or sanctions advice. Returns exposure dimensions, evidence gaps, a chokepoint-disruption watch, decision_readiness_score, and human_review_required: true.

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 live inside this runtime as productized service functions with their own schemas and HTTP/A2A profiles; four are also exposed as local stdio MCP tools, while Kazakhstan Market-Entry Readiness is currently HTTP/A2A-only. 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, agentic_interaction_trust, etc.) vendor-neutral, no transport, no marketplace
• MCP server — stdio server exposing 22 tools across the validation, product, and vertical worker 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; local shell documented in docs/deployment/a2a-adapter.md
• Cloudflare Worker baseline — deployment config under deploy/cloudflare-worker/; six live workers (general triage + the five vertical workers below)
• Vertical workers — productized service functions with their own schemas + HTTP/A2A profiles; Cloudflare deployments exist where configured. Currently shipped in the runtime: Middle Corridor Deal Risk Gate, CIS Secondary-Sanctions Exposure, Agentic Interaction Trust Gate, Gulf Maritime Exposure Gate, Kazakhstan Market-Entry Readiness Gate. Local stdio MCP exposes the first four vertical workers; Kazakhstan Market-Entry Readiness is HTTP/A2A-only.
• 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, check-memo-quality, memo-quality-bench, weekly-delta, 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 18 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

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/agentic-interaction/trust — Agentic Interaction Trust Gate (agentic-interaction-trust-request.schema.json)
• POST /v1/cis-secondary-sanctions/exposure — CIS Secondary-Sanctions Exposure triage (cis-secondary-sanctions-request.schema.json); live retrieval is active via the $0 Snapshot upstream (or set WATCHMAN_URL / OPENSANCTIONS_API_KEY), degrading gracefully to user-supplied evidence only if unavailable
• POST /v1/market-entry/readiness — Kazakhstan Market-Entry Readiness Gate (market-entry-readiness-request.schema.json); HTTP/A2A-only in the current release

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:1.1.1 .
docker run --rm -p 8080:8080 agenda-intelligence-md-api:1.1.1

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.

More CLI examples

agenda-intelligence bench examples/source-backed --strict --min-score 80
agenda-intelligence memo-quality-bench tests/fixtures/memo_quality --format json
agenda-intelligence weekly-delta-bench tests/fixtures/weekly_delta --format json
agenda-intelligence audit-claims examples/source-backed/eu-ai-act.audit.json --strict
agenda-intelligence weekly-delta examples/strategic-infrastructure-bankability/status.synthetic.md --format json
agenda-intelligence mcp-config --client cursor

Pinned-wheel install (instead of PyPI):

pip install https://github.com/vassiliylakhonin/agenda-intelligence-md/releases/download/v1.1.2/agenda_intelligence_md-1.1.2-py3-none-any.whl

Benchmark baseline

20 source-backed cases, reproduced with agenda-intelligence bench examples/source-backed/. The score below measures structural completeness — schema validity, evidence labeling, source-coverage diagnostics, and decision-readiness — not factual accuracy. A high score means a brief is well-formed and audit-ready, not that its claims are true in the world.

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

Benchmark your own output. The same harness scores your agent's briefs and compares them to this baseline — point bench at a directory of your <name>.brief.json cases. No LLM, no network: docs/benchmark-your-output.md.

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

AnalysisBank

AnalysisBank stores reusable reasoning memories, not factual state or source archives. memory-search retrieves only active, unexpired lessons by default; memory-lint validates lifecycle metadata, card schema conformance, and memory_index.json sync.

agenda-intelligence memory-lint
agenda-intelligence memory-search sanctions
agenda-intelligence memory-search-bench tests/fixtures/analysis_bank_retrieval/manifest.json
agenda-intelligence memory-applicability-bench tests/fixtures/analysis_bank_applicability/manifest.json

The retrieval bench checks representative queries against expected top lessons and forbidden top-N mistakes. The applicability bench checks positive and negative task contexts against each card's Apply when / Do not apply when boundary. analyze uses the same guarded path before adding a bounded REASONING MEMORY section to the system prompt and audit.reasoning_memory trace to the memo. These checks evaluate reasoning-memory routing discipline, not factual truth.

Release/ops checkpoint: docs/product/analysisbank-hardening-checkpoint.md. It records the before/after state, runtime path, operator checklist, quality gates, and stop rule for future AnalysisBank changes.

MCP

Stdio MCP server with 22 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)
├─ okf/                       # compact agent-readable knowledge bundle
├─ .well-known/               # agent card and Agentic Resource Discovery catalog source artifacts
├─ 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/, analysis-bank/ — 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