nova.mcp-bcp-counting
Enables Claude to guide developers through structured complexity estimation using BCP Plus methodology, counting 13 dimensions and persisting results for H/BCP tracking.
README
nova.mcp-bcp-counting
MCP Server for BCP Plus complexity counting — 10 functional + 3 non-functional dimensions.
Built on the BCP Plus methodology (CI&T, May 2026). Enables Claude to guide developers through structured complexity estimation during planning or coding sessions in VSCode, Claude Code, and Cursor.
What it does
When a developer asks Claude to count BCP for a story, Claude automatically:
- Assesses story maturity — warns if the story is too vague to count reliably
- Counts each dimension in isolation — 1 prompt per dimension, reducing cross-dimensional interference (per whitepaper recommendations)
- Calculates BCP Plus total — Functional BCP + NFR BCP + NFR Ratio + risk classification
- Persists the result — saves to
.bcp/counts.jsonin the project for H/BCP tracking
Quick start
Prerequisites
- Node.js ≥ 18
- One of: VSCode, Claude Code, or Cursor
Install
# Clone and build
git clone https://github.com/danielvm-git/nova.mcp-bcp-counting.git
cd nova.mcp-bcp-counting
npm install
npm run build
# Install for your platform
node dist/installer/install.js --platform vscode # VSCode
node dist/installer/install.js --platform claude-code # Claude Code
node dist/installer/install.js --platform cursor # Cursor
# Or interactive (asks which platform)
node dist/installer/install.js
Use
Open a story file or paste it in the chat, then ask Claude:
Count the BCP Plus for this story.
Decompose this epic into stories and estimate the BCP for each.
Show me the project BCP metrics so far.
The 13 dimensions
Functional (D01–D10) — inherited from BCP open-source release
| # | Dimension | CV% | Stability |
|---|---|---|---|
| D01 | Business Rules | 19.2% | Moderate |
| D02 | Interface Elements | 40.1% | High ⚠️ |
| D03 | Roles / Permissions | 33.2% | High ⚠️ |
| D04 | Solution Variabilities | 47.1% | Critical 🔴 |
| D05 | Boundaries | 44.6% | Critical 🔴 |
| D06 | Domain Entities | 38.6% | High ⚠️ |
| D07 | New Domain Entities | 24.8% | Moderate |
| D08 | Background Processes | 21.4% | Moderate |
| D09 | Notifications | 13.5% | Stable ✅ |
| D10 | Audits | 13.2% | Stable ✅ |
Non-functional (D11–D13) — BCP Plus exclusive
| # | Dimension | CV% | Stability |
|---|---|---|---|
| D11 | Quality Attributes | 22.5% | Moderate |
| D12 | Security & Compliance | 10.5% | Stable ✅ |
| D13 | UX & Accessibility | 9.8% | Most stable ✅ |
CV% = coefficient of variation from the 43-story empirical baseline (CI&T Flow team, April 2026).
D04 and D05 are flagged as LOW CONFIDENCE in all outputs — always validate with the team.
Fibonacci scale
| Label | Score |
|---|---|
| XS | 1 |
| S | 2 |
| M | 3 |
| L | 5 |
| XL | 8 |
NFR Ratio risk levels
| NFR Ratio | Risk | Action |
|---|---|---|
| 0–15% | Primarily functional | Standard delivery |
| 15–30% | Balanced | Monitor NFR during sprint |
| 30–50% | NFR-heavy 🟡 | Architectural spike recommended |
| >50% | Extreme 🔴 | Architectural review required |
Available MCP tools
| Tool | Description |
|---|---|
assess_story_maturity |
Score story maturity 1–5 and get expected CV% |
list_dimensions |
List all 13 dimensions with stability status |
get_dimension_guide |
Full sizing criteria + per-dimension prompt for a given D-id |
count_dimension |
Get counting context for a dimension (used by Claude per-dimension) |
calculate_bcp_plus |
Aggregate scores → Functional + NFR + NFR Ratio + risk |
decompose_epic |
Get guidance to split an epic into well-sized stories |
save_story_count |
Persist a count to .bcp/counts.json |
get_project_metrics |
Read all saved counts and project totals |
Story maturity and counting reliability
Story maturity is the strongest predictor of counting accuracy:
| Maturity | Expected CV | Recommendation |
|---|---|---|
| < 3 | ~39.5% | ⚠️ Do not count — refine first |
| = 3 | ~15.5% | Acceptable — flag open items |
| > 3 | ~12.2% | ✅ Ready to count |
A story reaches maturity 3+ when it has:
- A defined actor
- Clear acceptance criteria (Given/When/Then)
- Business value stated
- Key entities named
Project metrics
After counting stories, BCP data is saved locally in .bcp/counts.json:
{
"project": "my-project",
"summary": {
"story_count": 12,
"total_bcp": 148,
"avg_bcp_per_story": 12.3,
"avg_nfr_ratio_pct": 18.5
},
"stories": [...]
}
Use total_bcp + actual delivery hours to compute H/BCP (hours per BCP point) — the key productivity indicator for AI-augmented teams.
Platform config files generated
| Platform | File |
|---|---|
| VSCode | .vscode/mcp.json |
| Claude Code | via claude mcp add |
| Cursor | .cursor/mcp.json |
References
- BCP Plus Whitepaper — Daniel Vieira Magalhães, CI&T, May 2026
- ScopeCounting — related scope counting tooling
- BCP open-source release — CI&T & Itaú Unibanco, MIT License, May 2026
License
MIT — © CI&T
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.