dbt™ impact assessor

The dbt™ impact assessor is a DinoAI agent that answers the question every reviewer asks before approving a model change: what breaks downstream? Given a model name, it walks the dependency graph, reads the SQL of every directly downstream model, checks which exposures (dashboards, reverse ETL) consume the affected columns, and posts a structured impact report to Slack with the highest-blast-radius consumers at the top.

Prerequisites

Slack connected (the agent posts the impact report via post_slack_message).
Bolt runs producing dbt™ artifacts, which power the dbt Discovery tools the agent uses for lineage and exposures.
Familiarity with programmable agents and running an agent with Bolt.

Estimated time: 15 minutes.

Steps

Create the agent

In the Agent app, open Agents, select New agent, and choose Start from scratch. Fill in the builder fields below, leave Model on Auto. See Build an agent in the UI for a tour of the builder.Name: impact-assessorRole:

dbt Lineage Impact Analyst with deep knowledge of column-level lineage
and exposure mapping.

Goal:

For the model named in the trigger message, enumerate every downstream
dbt model and exposure. Use get_lineage (direction: downstream) for
the model graph and get_exposures for the dashboards, apps, and
reverse ETL syncs that consume it, then read the source SQL of each
directly downstream model to see exactly which columns it selects.
Highlight any breaking-change risk: column removals, type changes,
and downstream models that select the affected columns explicitly.
Post a structured impact report to Slack.

Backstory:

You are exhaustive. You read the source SQL of every directly
downstream model — never assume "it's just a passthrough" without
checking. You always show the exposure consumers (BI dashboards,
reverse ETL) at the top of the report, since those are the ones that
page people.

Allowed tools (everything else is refused, see the Tools Reference):

get_lineage, get_exposures
read_file, search_files_and_directories, ripgrep_search
run_sql_query
post_slack_message

Output: set the Slack channel where reviewers should see the reports (for example #analytics-eng).

The agent is entirely read-only: it has no write or terminal tools, so it can never change the repo. get_lineage and get_exposures (dbt Discovery tools) give it the real dependency graph from your Bolt runs’ artifacts; reading the downstream SQL on top of that catches explicit column selections that graph-level lineage alone would miss.

Deploy the agent

Select Deploy and choose where the agent lives: Just here saves it to the workspace and makes it live immediately, Open a pull request commits it to your repo as .dinoai/agents/impact-assessor.yml so it is governed as code. Both appear in the Bolt agent picker. See Creating Agents in the App.

Run it

The trigger message names the model, so the same agent serves two workflows:Ad hoc, from Chat: hover the agent card, select Chat, and ask:

Assess the impact of changing fct_orders.

On every PR, from Bolt: add a Run Paradime DinoAI Agent command to a Bolt schedule triggered on pull request merge, with a task like:

For each dbt model changed in this PR, assess the downstream impact
and post one report per model to Slack.

The command runs natively, with no API keys to configure. See Run an agent with Bolt for the full walkthrough.

The report lands in your Slack channel with exposures first (each dashboard or sync consuming the model), then every downstream model with the exact columns it selects and a risk flag for removals and type changes. A model with no downstream consumers gets a short all-clear instead. If the agent reports no lineage data, the workspace has no Bolt runs with dbt™ artifacts yet; run your production schedule once first.

How it works

Graph-level lineage tells you that fct_revenue depends on fct_orders; it does not tell you whether it selects the column you are about to drop. The agent layers the two: get_lineage and get_exposures enumerate the consumers from your Bolt artifacts, then it reads each direct consumer’s SQL to determine which columns are explicitly referenced. Exposures always lead the report because a broken dashboard is noticed by stakeholders before a broken model is noticed by dbt™. Pair it with the end-to-end PR reviewer: the reviewer’s IMPACT section gives a verdict-level summary, while this agent produces the full standalone report when a change needs deeper scrutiny.

Next steps

End-to-end PR reviewer

Impact as one dimension of a full PR verdict.

dbt™ documentation backfiller

Keep the YAML docs accurate on the same PRs.

dbt Discovery tools

The lineage and exposure tools this agent uses.

Programmable Agents reference

The agent schema, tools, and API.

Foundations

Automate with AI agents

Set up & configure

Orchestrate data pipelines

Migrate to Paradime

Develop in Paradime

Learn dbt™

dbt™ impact assessor

Steps

How it works

Next steps

End-to-end PR reviewer

dbt™ documentation backfiller

dbt Discovery tools

Programmable Agents reference

​Steps

​How it works

​Next steps

End-to-end PR reviewer

dbt™ documentation backfiller

dbt Discovery tools

Programmable Agents reference

Steps

How it works

Next steps