agentcompass config

Inspect merged configuration values and live component parameter schemas.

agentcompass config separates effective values from component field documentation:

agentcompass config COMMAND

config show

Print the merged effective configuration:

agentcompass config show [OPTIONS]

Without component selectors, it prints global runtime and execution sections. Add selectors to include the defaults for a benchmark, harness, and environment:

agentcompass config show \
  --benchmark swebench_verified \
  --harness mini_swe_agent \
  --env docker

Use repeatable --config values to inspect the same layering that a run will use, and choose JSON when another tool will consume the output:

agentcompass config show \
  --config configs/shared.yaml \
  --config configs/private.yaml \
  --benchmark swebench_verified \
  --format json

Redirect the YAML output when you want an editable starting point:

agentcompass config show \
  --benchmark swebench_verified \
  --harness mini_swe_agent \
  --env docker \
  > configs/swebench-verified-docker.yaml

config docs

Print the live typed fields, defaults, and descriptions for one component:

agentcompass config docs benchmark swebench_verified
agentcompass config docs harness mini_swe_agent
agentcompass config docs env docker

Use config show to answer “what value is effective?” and config docs to answer “what does this component accept?”. Model parameters are governed jointly by the selected harness, API protocol, and endpoint, so consult Configure a Model as well.

Precedence and Secrets

Explicit CLI and JSON parameter flags override matching configuration-file values. Recipes then adapt task metadata while preserving supported explicit user overrides. See Configuration Files and Precedence for the complete order.

Do not commit API keys or provider tokens in generated configuration. Keep credentials in environment variables or a private configuration file excluded from version control.