Skip to content

Discover

Tutorials, labs, and integrations

This section is the hands-on path: tutorials, labs, and integrations.

The left menu lists every tutorial, every lab family, and the integration pages. Use this index to pick a route. Open protocols (OpenAI, MCP, A2A, Agent Skills, OpenTelemetry) live under Open standards.

  • Tutorials

    From one agent to a governed experiment. Start with Tutorial 0.

    Tutorials index

  • Labs

    Packaged experiments: IoC motivation and protocol studies, plus the three paper labs shipped with this repository.

    Labs index

  • IoC integrations

    How Cognition Protocol skills go in, and how run records come out for SCOPE.

    IoC integrations

  • Third party

    Labs and resources published outside this repository, and how to point MAS-Lab at them.

    Third party

  • Open standards

    OpenAI Chat Completions, MCP, A2A, Agent Skills, and OpenTelemetry through AGNTCY's observe-sdk.

    Open standards

Tutorials

Do Tutorials 0 to 3 in order: install, then one agent, then a MAS, then experiments. Tutorials 4 to 6 and 8 to 14 each cover one capability.

# Tutorial What you learn
0 Environment setup Install, PATH, model endpoint, API key wiring
1 Build an agent Agent manifest, tools, skills, memory, CLI, traces
2 Orchestrate your MAS MAS manifest, delegation, topology overlays
3 Run an experiment Experiments, benchmarks, pipelines
4 MCP tools Local and remote tools through one contract
5 A2A agents Local and remote agents through one contract
6 Agent skills Reusable skills outside an agent's own instructions
8 Telemetry Native traces, live OTel, replay
9 KG & SCOPE Knowledge graph from traces
10 Evaluation metrics Scores, judges, reusable metrics
11 Sessions and recovery Checkpoint, fork, resume
12 Spawned subagents Bounded subagent templates
13 Control attach and debug Interrupt, persist, resume, debug scripts
14 Governance and HITL Policy and human approval

Labs

Family Question Where
IoC motivation How do multi-agent systems fail? Labs index
Cognition protocols Do ACCORD and CONCORD repair those failures, and at what cost? Labs index
Cognitive-challenge Do the failures show up on other systems too? Labs index
Paper labs Is the method itself sound? Labs index

Integrations

Path What it is
Cognition Protocols as skills ACCORD and CONCORD enter as initiator and receiver skills
Runs out to SCOPE Live telemetry or conversion from a recorded run
Knowledge graph locally Build the same graph SCOPE would, on your machine
Third party Labs and resources published outside this repository