References¶
This hub lists the reference pages for MAS Lab: YAML manifests, schemas, runtime contracts, and lab benchmarking. Unsure about a term? Check the glossary. Open protocols (OpenAI, MCP, A2A, Agent Skills, OpenTelemetry) live under Open standards, not here.
Day-to-day topics live under User guide: command-line index, Web UI, live dashboard.
CLI¶
Console scripts, flags, and the controller HTTP API.
| Topic | Reference |
|---|---|
mas-runtime |
mas-runtime.md |
mas-ctl |
mas-ctl.md · compile · overrides |
mas-lab |
mas-lab.md |
| Run logs | Observability |
| Controller HTTP API | Controller API |
| Live dashboard internals | live-reference.md |
How the surfaces fit together: command-line index (User guide).
Specifications¶
The YAML kinds and how they compose.
| Topic | Reference |
|---|---|
| Overview & composition | Manifest overview |
| Agent | agent.md · governance.md (spec.governance) · checkpoints · subagents |
| MAS & workflow | mas.md |
| Overlay | overlay.md |
| Flavour & environment | flavour.md, infra.md |
| Workspace file | config.yaml, user-config.md |
| Compiled defaults | What a minimal agent expands to |
| Tools | ToolContract · ToolServerRegistry · System tools |
| Schemas (all) | Schema index |
| Contracts | Contracts |
Runtime¶
The execution kernel, contracts, and plugins (mas-runtime, mas-ctl).
| Topic | Reference |
|---|---|
| Runtime manifests (Agent, MAS, overlay) | Manifest fields |
| Context window, summarization, pinned prompt parts, cache and token usage | Context management |
Reliability (retries, circuit, error_policy, logging) |
reliability.md |
LLM response cache (llm_cache infra middleware) |
llm-cache.md |
| Contracts & Mealy envelope | Contracts · runtime package docs |
| Thin waist (plugin vs catalog) | thin-waist.md |
| Snapshots vs checkpoints | snapshots.md |
| Kernel operations | kernel-primitives.md |
| Run logs | Observability (CLI) |
| Native event stream | Event model |
Lab¶
Experiments, datasets, pipelines, and analysis (mas-lab).
| Topic | Reference |
|---|---|
| Experiment manifest | experiment.md |
| Dataset | dataset.md |
| Pipeline steps | pipeline.md |
| Interactive lab demo | lab.md |
| Bench design & steps | Lab package docs |
| Library discovery | library-discovery.md |
Packages¶
| Topic | Reference |
|---|---|
| Package map | libraries.md |
| Generated package list | packages-reference.md |
| Generated plugin list | plugins-reference.md |
See also¶
- Getting started: install.
- User guide: day-to-day workflows.
- Open standards: OpenAI, MCP, A2A, Agent Skills, OpenTelemetry.
- Tutorials: hands-on, from agents to teams to experiments.
- Paper labs: reproduce Section 5 of the article.