---

name: analysis-experiments description: "Collect Mesa experiment data, run scenario sweeps, and use experimental actions, states, signals, and meta-agent analysis helpers." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Analysis Experiments

Use this sub-skill when a Mesa task is about experiment outputs, scenario orchestration, or experimental analysis helpers rather than basic model structure or visualization.

Use for

• DataCollector reporters, model / agent / agent-type DataFrames, and manual tables.
• Experimental mesa.experimental.data_collection datasets, registries, and recorders.
• Experimental Scenario, RunConfiguration, run_scenarios, stores, statuses, and failure diagnostics.
• Experimental timed Action lifecycles, continuous states, thresholds, reactive signals, and batching.
• Experimental MetaAgent grouping utilities for higher-order agent analysis.

Start here

• API reference for signatures, import paths, return shapes, and stability notes.
• Workflows for copyable collection, recorder, scenario, signal/state/action, and meta-agent recipes.
• Troubleshooting for reporter validation, recorder setup, scenario failure origins, pickling, and experimental API pitfalls.
• DataCollector smoke script for a safe installed-package check that emits JSON summaries.
• Scenario smoke script for a safe installed-package check of Scenario, RunConfiguration, run_scenarios, and failure recording.

Route elsewhere

• Core Model, Agent, AgentSet, activation, RNG, and event scheduling basics: model-core
• Spatial placement, movement, grids, networks, cells, or continuous-space geometry: spaces
• Solara dashboards, browser visualization, plotting widgets, charts, or rendering: visualization

All APIs under mesa.experimental are experimental and may change between Mesa releases. Treat this sub-skill as the self-contained operating reference; no external repository files are required at runtime.