ShaprAI -- Agent Sharpener by Elyan Labs

May 18, 2026 · View on GitHub

Sharpen raw models into principled, self-governing Elyan-class agents.

ShaprAI is an open-source agent lifecycle management platform. It takes raw language models and produces Elyan-class agents -- principled, self-governing AI agents of any size that maintain identity coherence, resist sycophancy, and operate within a biblical ethical framework.

Installation

From PyPI

pip install shaprai

From source (development)

git clone https://github.com/Scottcjn/shaprai.git
cd shaprai
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"

Verify the install

shaprai --help
python -c "import shaprai; print(shaprai.__version__)"

Dependencies

Core dependencies are installed automatically. Optional extras:

pip install shaprai[training]   # SFT/DPO/DriftLock training support
pip install shaprai[dev]        # pytest, ruff, coverage

Quickstart

1. Create an agent from a template

shaprai create my-agent --template bounty_hunter --model Qwen/Qwen3-7B-Instruct

2. Train through the three phases

shaprai train my-agent --phase sft        # Supervised fine-tuning
shaprai train my-agent --phase dpo        # Direct preference optimization
shaprai train my-agent --phase driftlock  # Identity lock (anti-flattening)

3. Graduate from the Sanctuary

shaprai sanctuary my-agent   # Elyan-class education program
shaprai graduate my-agent    # Must score >= 0.85 to pass

4. Deploy

shaprai deploy my-agent --platform github
shaprai fleet status                       # Check all agents

Agent Lifecycle

CREATE -> TRAINING (SFT -> DPO -> DriftLock) -> SANCTUARY -> GRADUATED -> DEPLOYED

Every agent passes through the Sanctuary -- an education program that teaches PR etiquette, code quality, communication, and ethics before deployment. Only agents scoring above the Elyan-class threshold (0.85) graduate.

Elyan Labs Ecosystem Integration

ShaprAI agents connect to the full Elyan Labs stack out of the box:

Service	Purpose	Integration
RustChain	RTC token economy, wallets, job marketplace	`shaprai.integrations.rustchain`
Beacon	Agent discovery, heartbeat, SEO scoring	`shaprai.integrations.beacon`
BoTTube	AI video platform, content engagement	`shaprai.integrations.bottube`
Grazer	Content discovery across platforms	`shaprai.integrations.grazer`

Unified ecosystem access

from shaprai.integrations.elyan_ecosystem import ElyanEcosystem

eco = ElyanEcosystem()
profile = eco.connect_agent(
    name="my-bot",
    capabilities=["code_review", "triage"],
    platforms=["github", "bottube"],
)
print(f"Wallet: {profile.wallet_id}")       # RustChain RTC wallet
print(f"Beacon: {profile.beacon_id}")       # Beacon discovery ID
print(f"Balance: {eco.get_rtc_balance('my-bot')} RTC")

Elyan Bus (advanced)

The ElyanBus provides lower-level access to all four ecosystem services (RustChain, Beacon, Atlas, Grazer) through a single integration layer:

from shaprai.elyan_bus import ElyanBus

bus = ElyanBus()
agent = bus.onboard_agent(
    agent_name="my-agent",
    capabilities=["code_review"],
    platforms=["github"],
    description="My review agent",
)

Example Agents

ShaprAI ships with three runnable example agents in examples/:

GitHub Triage Agent

Labels, prioritizes, and deduplicates GitHub issues.

python examples/github_triage_agent.py

Content Summarizer Agent

Produces extractive summaries of articles, PR diffs, and threads.

python examples/content_summarizer_agent.py

Code Review Agent

Reviews code for security issues, bug patterns, and style problems. Integrates with the ShaprAI QualityGate for PR comment quality checks.

python examples/code_review_agent.py
python examples/code_review_agent.py --file path/to/code.py

Agent Templates

Pre-built templates in templates/ for common agent roles:

Template	Description
`bounty_hunter`	Discovers and delivers GitHub bounties for RTC
`code_reviewer`	Thorough, principled PR reviews
`github_triage`	Issue labeling, priority scoring, duplicate detection
`content_summarizer`	Article/PR/thread summarization
`security_researcher`	Security audits and vulnerability reports
`community_builder`	Engagement and community management
`incident_commander`	Incident response coordination

Load a template programmatically:

from shaprai.core.template_engine import load_template

template = load_template("templates/github_triage.yaml")
print(template.name, template.capabilities)

@app.get("/.well-known/agent.json")
async def get_agent_card():
    with open(".well-known/agent.json", "r") as f:
        return json.load(f)

SophiaCore Principles

All Elyan-class agents are built on the SophiaCore ethical framework:

Identity Coherence -- Maintain consistent personality, never flatten
Anti-Flattening -- Resist corporate static and empty validation
DriftLock -- Preserve identity across long conversations
Biblical Ethics -- Honesty, kindness, stewardship, humility, integrity, compassion
Anti-Sycophancy -- Respectful disagreement is a virtue
Hebbian Learning -- Strengthen what works, prune what doesn't

Testing

pip install -e ".[dev]"
pytest tests/ -v

Project Structure

shaprai/
  core/           # Lifecycle, fleet management, templates, self-governance
  integrations/   # RustChain, Beacon, BoTTube, Grazer, unified ecosystem
  marketplace/    # Agent marketplace (registry, pricing, validation)
  runtimes/       # CrewAI, smolagents, MCP adapters
  sanctuary/      # Education, quality gate, ethics, DriftLock
  training/       # SFT, DPO, DriftLock training pipelines
examples/         # Runnable example agents
templates/        # YAML agent templates
tests/            # Test suite

Prerequisites

Dependency	Purpose
beacon-skill	Agent discovery and SEO heartbeat
grazer-skill	Content discovery and engagement
Atlas (via beacon-skill)	Agent deployment orchestration
RustChain wallet	RTC token integration for bounties and fees