Reputation Portability: The Missing Infrastructure for Economic Mobility

#africa #labor #python #mcp

A skilled plumber in Chicago can move to Los Angeles and immediately find work. Her Yelp reviews, contractor license, and Angie's List history travel with her. Strangers trust her before she's done a single job.

A skilled mason in Mombasa cannot take her reputation to Nairobi. Trust is tribal and local. Hiring defaults to personal networks. The most talented people outside existing networks are invisible.

This is not a cultural problem. It's an infrastructure problem.

What Reputation Portability Infrastructure Looks Like

In mature economies, the infrastructure exists in fragments:

Uber/Lyft: Star ratings that follow a driver across cities
LinkedIn: Endorsements and work history that cross borders
BBB (Better Business Bureau): Verified complaint and resolution history
Angie's List / Thumbtack: Job completion ratings for home services
Upwork JSS: Job Success Score based on completed contracts
Professional licensing boards: Verifiable credentials (doctor, electrician, contractor)

Each of these solves the same problem: how do you compress trust between strangers at scale?

sifa-mcp (Sifa = reputation/praise in Swahili) is a portable reputation MCP server that implements this infrastructure for East Africa's informal economy.

The Trust Score Architecture

from sifa_mcp import create_worker_profile, add_service_record, get_reputation_profile

# Create a portable profile
profile = create_worker_profile(
    full_name="Amina Odhiambo",
    phone="0712345678",
    primary_skill="construction",
    sub_skills="masonry,tiling,plastering",
    location="Mombasa",
    years_experience=8
)
# → profile_id: "SFA-A3F2B19C4E1D"

# Add a completed job
add_service_record(
    profile_id="SFA-A3F2B19C4E1D",
    client_name="ABC Contractors",
    client_phone="0722000000",
    service_type="Kitchen tiling",
    duration_days=5,
    payment_kes=15000,
    client_rating=5
)

# Check trust for a new employer
profile_data = get_reputation_profile(
    profile_id="SFA-A3F2B19C4E1D",
    requester_context="employer"
)
# → trust_tier: "ESTABLISHED"
# → recommendation: "Good track record. Suitable for most engagements."

The Score Decomposition

The trust score (0–100) is built from four signals:

Signal	Max Points	Rationale
Job volume	30	Demonstrates activity, not just registration
Average quality	35	Client-rated — heaviest weight
Verified records	20	Client OTP confirmation prevents fabrication
Tenure	15	Long history = stability signal

This maps directly to how humans assess trust: frequency + quality + verification + history.

The Production Path

sifa-mcp is DEMO infrastructure. The in-memory profile store needs to be backed by a persistent database. More critically, production requires identity verification tied to Kenya's Huduma Number / NIIMS system — to prevent workers from fabricating verified records.

The regulatory and identity infrastructure for this exists in Kenya. The coordination layer (portable reputation MCP) does not. That's the gap sifa-mcp fills — and the conversation it starts.

Why NITA Alignment Matters

The skill categories in sifa-mcp are aligned with Kenya's National Industrial Training Authority trade classifications. This isn't accidental — NITA certification is the existing institutional trust anchor for skilled tradespeople. A sifa-mcp score that references NITA-certified training becomes part of a verifiable credential chain.