Context Mapping

Context Mapping connects your scenario data to the variables used in your test flow. Each mapping entry takes a field from the scenario and makes its value available as a {{VARIABLE}} in your steps, request bodies, and assertion rubrics.

How to add a mapping

In Test Builder, go to the Context Mapping tab. Click Add mapping, then:

1. Pick the context field you want to map (use the field tree to select nested fields)
2. Choose a variable name (e.g. USER_ID)
3. Choose the mapping type
4. Click Add

The variable is then available as {{USER_ID}} anywhere in the test.

Mapping types

direct

The field value is used as-is. Best for simple scalar values like strings and numbers.

field: id  →  variable: USER_ID  →  {{USER_ID}} = "5550000001"

json

The field value is serialized as JSON. Useful for passing a structured object or array as a variable.

field: watch_history  →  variable: WATCH_HISTORY  →  {{WATCH_HISTORY}} = "[{\"title\":\"Barbie\",...}]"

summary

The field value is formatted as a human-readable summary. Useful for injecting array data into an LLM rubric where you want readable text rather than raw JSON.

field: friend_activity  →  variable: FRIEND_ACTIVITY  →  {{FRIEND_ACTIVITY}} = "Barbie, Oppenheimer, ..."

jmespath

A JMESPath expression is evaluated against the scenario data to extract a specific subset. This is the most powerful option — it lets you filter, pluck, and transform scenario data.

field: watch_history[*].title
→  variable: already_watched
→  {{already_watched}} = "Barbie, Oppenheimer, The Batman"

Common JMESPath patterns:

Using mapped variables in rubrics

Once mapped, variables are available in assertion rubrics:

Must not recommend any title from: {{already_watched}}

Prioritize titles from: {{friend_recent_likes}}

This makes your evaluations scenario-specific — the rubric references actual data from each user's scenario rather than generic criteria.