AI Delivery Coordination

Companion orchestration engine for ai-delivery-kit. Exposes an MCP server for tool-based integration; the kit skill layer must not import this package — use MCP only.

Quick start

python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest tests/unit --ignore=tests/unit/test_task10_crew.py --ignore=tests/unit/test_crew_bridge.py

Related MCP server: LLM Bus

ai-delivery-kit bridge

orchestration/skill_bridge.py loads .ai-delivery/requirements/<req-id>/status.json into a pipeline view and can flush execution-segment statuses (in_dev / merged / blocked_*) back before reconcile runs.

MCP loop tools: start_loop, stop_loop, loop_status, stall_report, intervene_loop.

Operator guide in ai-delivery-kit: .agents/skills/ai-delivery-orchestrator/references/coordination-mcp-bridge.md.

Layout contract

Reads the same project-binding.json → layout paths as ai-delivery-kit (config/paths.py). No skill-layer Python imports.

Repository layout

Vendored in ai-delivery-kit as the ai-delivery-coordination/ git submodule. Clone the kit with:

git clone --recurse-submodules https://github.com/s-charvin/ai-delivery-kit.git