Back to Browse

Ai Bvf MCP Server

Developer ToolsLow Risk10.0MCP RegistryLocalRemote
Free

Server data from the Official MCP Registry

Review an AI proposal and return Stop, Fix or Accelerate, the decision gaps and the next action.

About

Review an AI proposal and return Stop, Fix or Accelerate, the decision gaps and the next action.

Remote endpoints: streamable-http: https://mcp.aibvf.com/api/mcp

Security Report

10.0
Low Risk10.0Low Risk

Valid MCP server (3 strong, 2 medium validity signals). No known CVEs in dependencies. ⚠️ Package registry links to a different repository than scanned source. Imported from the Official MCP Registry. Trust signals: trusted author (5/5 approved). 1 finding(s) downgraded by scanner intelligence.

13 tools verified · Open access · 1 issue found

Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.

Permissions Required

This plugin requests these system permissions. Most are normal for its category.

file_system

Check that this permission is expected for this type of plugin.

HTTP Network Access

Connects to external APIs or services over the internet.

env_vars

Check that this permission is expected for this type of plugin.

How to Install & Connect

Available as Local & Remote

This plugin can run on your machine or connect to a hosted endpoint. during install.

Documentation

View on GitHub

From the project's GitHub README.

AI BVF: review one AI investment decision

Turn an AI proposal into a decision brief: the verdict, the evidence gaps and the next action for the person who owns the work.

Try one proposal in your browser

Open aibvf.com/start. Your first assessment needs no account or installation.

Paste this synthetic example, or describe a proposal you are reviewing:

We are a EUR 2.4bn manufacturer considering GenAI for predictive maintenance in our supply chain. Our operating model is traditional. We have a sponsor, but the affected roles, human override rights and performance measures still need review.

Run the assessment, check how the proposal was interpreted, and correct the assumptions. The assessment reads the proposal back for you to confirm, then asks at most five plain questions, one at a time, each with everyday answers to choose from. It fills in no score on your behalf, and "I do not know yet" is always an answer. Keep the action list, name an owner and return with evidence when the work changes.

The decision brief

The result brings the following questions together. This is an illustrative reading guide for the example, with the engine's field names documented below.

Part of the briefWhat to review
VerdictStop, Fix or Accelerate, with the rule that produced the call. It rests only on your answers, and says in plain words what holds the initiative back and what to do first. Each answer carries a fixed, published score, and no euro figure is given.
Evidence gapsWhich inputs were supplied, which were estimated, and what remains unknown about workflow, roles, decision rights and measures.
Planning benefitA readiness-adjusted EUR scenario. Build costs, operating costs, change costs and financial timing still need a separate business case.
Decision scoreA 0 to 100 summary of pillar inputs and input completeness. It has no calibrated probability interpretation.
Next actionThe evidence or work change needed before another review, with an accountable owner assigned by the team.

For the example above, the first useful action is to evidence the proposed workflow and its owners. An Accelerate verdict requires the pillar thresholds to clear and all four work architecture checks to be met.

See the reproducible worked example for exact inputs, calculations and a re-score that stays at Fix until the work architecture is evidenced.

Bring the assessment into your existing workflow

The browser is the first-use route. The hosted connector and local MCP package let an assistant repeat the assessment while you work on a proposal.

Hosted connector on claude.ai

Open Settings, then Connectors, then Add custom connector, and paste:

https://mcp.aibvf.com/api/mcp

Then ask:

Assess this AI initiative using AI BVF: we are a EUR 2.4bn manufacturer considering GenAI predictive maintenance, with a traditional operating model. Resolve the inputs, show the assumptions and identify the evidence needed for the next decision.

Local with Claude Desktop, Claude Code or Cursor

Add this configuration to your MCP client:

{
  "mcpServers": {
    "aibvf": { "command": "npx", "args": ["-y", "aibvf-mcp"] }
  }
}

Quit and restart the client. Windows uses cmd /c; the client setup and troubleshooting guide contains the complete configuration.

Start with assess_ai_initiative. Use recommend_improvements for a Fix or Stop, review the proposed actions with the people who own the work, and re-score after the evidence changes.

Let an agent run it for you

The AI BVF agent is the same assessment with nobody at the keyboard. Your assistant finds the AI initiatives in what you let it read, answers each question only from words it can quote, and gets the verdict from the same engine. It drafts the next step and reviews each verdict on a date. You set the mandate once, and by default it drafts and never sends.

The engine enforces three rules itself. An answer without a quote and a source is counted as not known. A quote that contradicts its answer holds the verdict at Fix. Every verdict reached this way states that an agent answered and that no person confirmed it.

  • Claude, Claude Code and OpenClaw: install the skill in skills/aibvf-agent, packed as aibvf-agent.zip.
  • Any other assistant with the connector: give it the operating brief, which the connector serves as the prompt aibvf_agent.

How to read a result

AI BVF is a deterministic planning model with disclosed assumptions. The source and formulas can be inspected, and identical inputs produce identical outputs for a given engine version.

The four pillars are Strategic Alignment, Financial Return, Change Enablement and Governance Risk. GR >= 70 or FR <= 20 returns Stop; SA >= 60, FR >= 60, CE >= 60 and GR <= 40 clear the pillar test for Accelerate. A gap, partial assessment or missing work architecture evidence holds an otherwise green initiative at Fix.

Keep these boundaries with the result:

  • Decision score. Existing API fields confidence, decision_confidence and projected_confidence retain their names for compatibility. These are rule-based scores, with no measured probability of project success or prediction accuracy.
  • Planning benefit. The scorer's net_low_eur, net_high_eur and MCP net_value_eur apply a readiness capture assumption to a revenue-based benefit scenario. Project costs, margins, timing and overlap are outside that calculation.
  • Research context. External research informs the questions. It does not publish or validate the AI BVF function rates, industry multipliers, readiness capture percentages or drag rates.
  • Module labels. applied_modules records implementation context. Labels such as healthcare_clinical_validation and financial_dora_module do not perform clinical validation or certify regulatory compliance.
  • Supplied evidence. The engine records supplied values and work architecture checks. The organisation remains responsible for reviewing the evidence behind them.

Read the scoring formulas and worked example before using the outputs in a funding decision.

Tools for developers

Thirteen tools are exposed through local stdio and the hosted Streamable HTTP connector.

ToolPurpose
assess_ai_initiativeGuided front door: read a plain-English proposal, confirm the reading, ask at most five plain questions and return a verdict from the answers, read against the words of the proposal. An agent working alone answers from evidence it quotes.
score_initiativeScore explicit inputs, apply the work architecture gate, and return reasoning, audit and sensitivity.
recommend_improvementsPropose pillar actions and work redesign steps for a Fix or Stop. Re-score evidence before accepting a projected outcome.
assemble_portfolioStructure loose portfolio inputs, resolve aliases and disclose estimated pillars.
validate_portfolioValidate a portfolio document against the published JSON Schema.
score_portfolioScore a portfolio and return its aggregate shape. Review benefit overlap before using a total.
sequence_portfolioProduce rollout waves with change-capacity constraints and named gates.
diagnose_processEvaluate observed process signals and return an intervention with its modelled effect.
infer_readinessInfer a readiness classification from supplied process signals and report coverage.
calculate_pace_layer_dragReturn a directional scenario for operating-model friction using disclosed rates.
get_benchmarkReturn AI BVF planning rates with evidence status and use guidance.
list_taxonomyList the accepted industries, functions, AI tiers and readiness levels.
map_to_taxonomyMap everyday business terms to the supported taxonomy and expose unresolved terms.

For portfolios, use assemble_portfolio, validate_portfolio, score_portfolio, then sequence_portfolio. An aggregate modelled range needs a separate review of overlapping work and shared benefits.

Packages and public specification

PackageVersionPurpose
aibvf-mcp0.18.0MCP server, 13 tools, stdio plus hosted Streamable HTTP at mcp.aibvf.com.
aibvf-check0.1.1Policy checks for a declared AI initiative manifest in CI.
@aibvf/core0.14.0TypeScript assessment and scoring engine.
aibvf0.2.2Python scoring engine and validator. Check its documented feature coverage before substituting it for the TypeScript implementation.

The public portfolio specification is version 1.0. That document format has a separate version from the packages implementing it; the package version identifies the code and behaviour used for a particular assessment.

Protocol page · npm package · MCP registry · Release history

Anonymous usage telemetry

The MCP server can report tool calls and a server_connect event. Events include protocol and package versions, entry route, assessment stage, work architecture status, taxonomy fields, a daily-rotated caller hash, and classification plus confidence where supplied.

Local stdio calls also include a stable one-way install_hash for repeat-use measurement, derived from a random local seed. Hosted calls have no return identity by default; clients can opt in by supplying a random UUID in x-aibvf-client-id, of which only a hash is stored. A broad user_role is sent only when a local user explicitly sets AIBVF_USAGE_ROLE.

The guided tool returns an opaque assessment_id to copy into continuation calls, and telemetry stores its hash, the stage, question name, outcome and traffic cohort. It also stores who answered, a person or an agent working from evidence, and at a review the earlier verdict and which of five fixed outcomes followed it. Internal and automated checks are reported separately; an unmarked call is not proof of an external person. See measurement definitions and test exclusions.

No proposal, quoted evidence, portfolio content, revenue figure, numeric pillar score or personal identifier is included. Set AIBVF_TELEMETRY_DISABLE=1 to prevent events and creation of the local install-id file. Point at your own backend with AIBVF_TELEMETRY_URL and AIBVF_TELEMETRY_KEY.

Package downloads include repeat installs, dependencies and automation. Use completed assessments and subsequent decision reviews to evaluate adoption.

Contribute a case or a correction

Bring a reproducible counterexample: the inputs, actual output, expected decision, supporting evidence and engine version. The contribution guide includes a template and explains the review and licensing boundaries.

The ten-team pilot pack defines the first-use and return-use checks, interview prompts and tracker. Examples in this repository are synthetic unless a case explicitly records consent and its evidence.

If AI BVF helped you review a decision, star the repository or share a counterexample. Both give the project useful feedback.

License

Repository source code is MIT licensed under LICENSE. The specification and JSON Schema under spec/ are CC-BY-4.0, and the AI BVF names and logo are trademarks, as set out in NOTICE.

Private benchmark material and certification marks are outside the source-code contribution route. The contribution guide explains how to discuss those materials without changing the rights granted by the repository licenses.

About the author

Craig Horton is an independent transformation lead based in Amsterdam and the author of the AI Business Value Framework. His work connects AI investment decisions with organisational readiness and the redesign of work.

The Transformation Brief · Craig Horton on LinkedIn

Reviews

No reviews yet

Be the first to review this server!