# EMBERBench documentation

Reviewed public task, scoring and development contracts.

- [Overview & setup](https://anantam-labs.vercel.app/docs/source/quickstart.md): The open task, public development results and first steps.
- [Command line](https://anantam-labs.vercel.app/docs/source/cli.md): Public generation, evaluation, replay and scoring commands.
- [Model interface](https://anantam-labs.vercel.app/docs/source/model-interface.md): Observations, actions and the decision-model contract.
- [Published site data](https://anantam-labs.vercel.app/docs/source/site-data.md): How this site is generated and what its snapshot establishes.
- [Simulation contract](https://anantam-labs.vercel.app/docs/source/simulation-contract.md): Movement, collisions, exits and fire semantics.
- [Scenario schema](https://anantam-labs.vercel.app/docs/source/scenario-schema.md): Public scenario fields and validation rules.
- [Public scenarios](https://anantam-labs.vercel.app/docs/source/scenario-suite.md): The diagnostic scenarios available for training and development.
- [Reference agents](https://anantam-labs.vercel.app/docs/source/baselines.md): Public hand-coded agents and planning references.
- [Development generator](https://anantam-labs.vercel.app/docs/source/development-generator.md): Open training worlds, separate from the official distribution.
- [EMBER scoring](https://anantam-labs.vercel.app/docs/source/scoring.md): The full formula, weights, gates, aggregation and failure handling.
- [Episode metrics](https://anantam-labs.vercel.app/docs/source/metrics.md): Raw outcomes and authenticated measurements.
- [Crowd diagnostics](https://anantam-labs.vercel.app/docs/source/crowd-diagnostics.md): Observed congestion and throughput signals; no crowd score.
- [Running evaluations](https://anantam-labs.vercel.app/docs/source/evaluation.md): Public paired schedules, records and result handling.
- [Frozen environments](https://anantam-labs.vercel.app/docs/source/environment-pack.md): Public pack identity, digests and matched fire worlds.
- [Statistical analysis](https://anantam-labs.vercel.app/docs/source/statistics.md): Intervals, resampling and paired comparisons.
- [Fairness contract](https://anantam-labs.vercel.app/docs/source/v1-fairness-spec.md): Model-visible information and simulator fairness.
- [Evaluation boundary](https://anantam-labs.vercel.app/docs/source/evaluation-boundary.md): What this site publishes, and how it differs from the official test.
- [Replay verification](https://anantam-labs.vercel.app/docs/source/replay-verification.md): Public verification rules and safe artifact publication.
- [Versions & identity](https://anantam-labs.vercel.app/docs/source/versioning.md): Public contract versions and what they identify.