> ## Documentation Index
> Fetch the complete documentation index at: https://docs.merionlabs.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Form an evaluation

> Convert governed work episodes into an independently approved and audited evaluation task.

# Form an evaluation

Merion treats evaluation engineering as a governed compilation process—not as automatic trace conversion.

```text theme={null}
reviewed episodes
  → cohort analysis
  → task qualification
  → four-part proposal
  → domain-expert interview
  → independent approvals
  → EvalTask build
  → evidence audit
  → ExperimentSpec link
```

## 1. Register evidence governance

Every episode must reference an active data-use policy that permits evaluation. Register the policy before analysis:

```bash theme={null}
merion evidence policy-put policy.json --workspace .merion
```

## 2. Analyze a cohort

The cohort records the sampling frame, inclusion and exclusion rules, schema inventory, coverage gaps, and proposed clusters.

```bash theme={null}
merion task cohort cohort-input.json --workspace .merion --json
```

## 3. Qualify a task candidate

Qualification explains economic importance, recurrence, capability uncertainty, representativeness, reproducibility, verifiability, non-triviality, rights/privacy, and outcome quality.

```bash theme={null}
merion task qualify qualification-input.json --workspace .merion --json
```

## 4. Draft four independent specifications

An evaluation proposal contains separately digest-pinned specifications:

* `TaskSpec`: realistic instruction, initial state, allowed actions, and expected end state.
* `HarnessSpec`: prompt, tools, permissions, budgets, and stop rules.
* `EvalEnvironmentSpec`: reset behavior, fidelity evidence, omissions, dependencies, and secrets boundary.
* `VerifierSpec`: capability axes, checks, calibration evidence, exploit hypotheses, and disagreement policy.

```bash theme={null}
merion task propose proposal-input.json --workspace .merion --json
merion task interview proposal.example --workspace .merion --json
```

## 5. Approve and audit

Each current specification digest requires an independent human approval. Changing a specification invalidates its approval.

After approval:

```bash theme={null}
merion task build proposal.example build-input.json --workspace .merion --json
merion task audit eval-task.example audit-input.json --workspace .merion --json
```

The audit checks resetability, data rights, exposure compatibility, solution leakage, environment fidelity, verifier independence, reproducibility, and human solvability.

## 6. Pin the experiment

```bash theme={null}
merion task experiment-link eval-task.example experiment.json --workspace .merion
```

This validates the exact task, environment, harness, and verifier boundary. It does not execute trials or train a model yet.