Scenario 38: Create a layer from a skill package
Use this when you want to configure a reusable layer from a remote skill package like dbt-labs/dbt-agent-skills without applying it to a project yet.
Frequency: Common · Status: Shipped
Use this when you want to configure a reusable layer from a remote skill
package like dbt-labs/dbt-agent-skills without applying it to a project yet.
Typical commands:
# Configure only (no hub install, no layer apply)
harnesstap layer create dbt-expert \
--from dbt-labs/dbt-agent-skills \
--all --exclude-category dbt-migration \
-d "dbt analytics engineering baseline" -y
# Interactive skill picker (migration skills unchecked by default)
harnesstap layer create dbt-expert --from dbt-labs/dbt-agent-skills
# Merge new skills into an existing layer
harnesstap layer create dbt-expert \
--from dbt-labs/dbt-agent-skills \
--skill running-dbt-commands \
--on-conflict merge -y
harnesstap layer show dbt-expertlayer create --from imports the package into the HarnessTap library and
attaches namespaced skill refs to the layer. Use layer apply when you are ready
to materialize the layer to a project. Use --install only when you also want
skills copied or symlinked into hub paths immediately.
See Scenario 35 for the lower-level add command
and Scenario 5 for manual layer curation.
Scenario 37: Publish a profile layer to the catalog
Use this when you want teammates to discover and install a switchable global preset from HarnessTap Cloud. Published profiles are ordinary published layers...
Scenario 39: MCP auth, environments, and account switching
Use this when you need to switch Slack workspaces, API tokens, or deployment targets across profiles/layers — or when OAuth MCP servers do not respond after...