Universal Software Development Control Plane

LinkedIn: https://www.linkedin.com/in/levkantorovich

A clean-room control plane for AI-assisted software development.

This repository demonstrates how agentic coding workflows can be made more reliable, reviewable, and production-oriented through structured tasks, procedural skills, quality gates, governance rules, explicit human handoff, and human review boundaries.

Portfolio purpose

This repository is intended as a portfolio artifact for roles involving AI-assisted engineering, agentic development workflows, AI infrastructure, software engineering productivity, platform engineering, systems engineering, security-minded automation, and code quality.

The goal is not maximum agent autonomy. The goal is AI-assisted engineering that is safer, smaller, more deterministic, easier to review, easier to test, and easier for another engineer to continue.

Current skills

• repository-safety
• task-execution
• quality-gates
• agent-handoff
• context-management
• kiss-engineering
• algorithmic-efficiency
• code-quality-comments

---

Original project documentation

A clean-room, reusable control plane for AI-assisted software development.

The repository provides a deterministic, human-controlled workflow around AI coding tools:

• explicit project and repository boundaries
• architecture before implementation
• large capability milestones rather than artificial micro-tasks
• separate architect, product, implementation, validation, security, and reality-check responsibilities
• generated task handoffs
• dry-run manifests
• test and evidence gates
• explicit non-goals and stop conditions
• truth-preserving completion reports

What this repository is

This is a generic software-development governance and orchestration layer. It is designed to sit beside a target software repository rather than contain the target product itself.

What this repository is not

• It is not a copy of any private production repository.
• It contains no arbitrage implementation, trading logic, credentials, or proprietary employer code.
• It does not automatically execute AI agents.
• It does not claim that passing unit tests proves production readiness.
• It does not require LangChain or LangGraph.

Public-review purpose

This repository is intentionally suitable for technical review. Reviewers can inspect:

1. docs/ARCHITECTURE.md
2. docs/GOVERNANCE.md
3. docs/REVIEWER_GUIDE.md
4. agents/
5. schemas/
6. projects/example_project/
7. tests/

The control plane can be linked from a target repository through a CONTROL_PLANE.md file.

Quick start

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[dev]'
pytest -q

sdcp validate-project projects/example_project/project.yaml
sdcp validate-task projects/example_project/tasks/EX-001.yaml
sdcp prepare-run \
  --project projects/example_project/project.yaml \
  --task projects/example_project/tasks/EX-001.yaml

Design principles

• Human approval remains explicit.
• Every task identifies what it proves and what it does not prove.
• Claims require evidence.
• Missing evidence remains missing evidence.
• Agents may not widen scope silently.
• Stop conditions are first-class.
• A task should unlock a reviewable capability, not merely add wrappers, schemas, or documentation.
• Product code stays in the target repository.
• Generated artifacts stay reproducible and inspectable.