Skip to Content

rag-forge assess

Run RAG Maturity Model assessment

Synopsis

rag-forge assess [options]

Description

assess runs a one-shot RMM (RAG Maturity Model) assessment against your pipeline configuration and produces a maturity level badge (0–5) with a pass/fail breakdown for each criterion.

Unlike audit, which requires telemetry or golden set data and makes LLM scoring calls, assess inspects your pipeline configuration and optional audit report to determine which maturity criteria are satisfied. It is fast, cheap, and suitable for running in CI to gate deployments on a minimum RMM level.

The command delegates to the Python rag_forge_evaluator.cli module. Results are printed to the terminal with a structured per-criterion breakdown. Each criterion shows whether it passed and which individual checks contributed to that verdict.

Options

FlagDefaultDescription
--config <json>Pipeline config as a JSON string
--audit-report <file>Path to a previously generated audit JSON report

Examples

Basic usage

rag-forge assess

Assess with an explicit config JSON

rag-forge assess --config '{"embedding_model":"text-embedding-3-small","chunk_size":512}'

Incorporate a prior audit report

rag-forge assess --audit-report ./reports/audit.json

Use in CI to check minimum maturity

rag-forge assess --config "$(cat rag-forge.config.json)" --audit-report ./reports/latest.json