@revura/native

Native gateway and tool runtime for Revura.

What it is

@revura/native is the strict agent/tool boundary for Revura.

Default stance:

• operators should start with the CLI on one machine
• agent/runtime integrations should start with revura-native-gateway over stdio
• loopback HTTP is available, but it is an advanced opt-in transport
• broader runtime/admin/automation surfaces remain implemented, but they are not the default product story

Default native surface

The default native bundle is the v1 accounting loop:

• post_ingestion_candidate
• post_ingestion_batch
• list_ingestion_reviews
• decide_ingestion_review
• reconcile
• import_reconciliation_statement
• generate_report
• dashboard_summary
• validate_close_controls
• close_period
• verify_chain

Everything else should be treated as explicit opt-in via REVURA_NATIVE_TOOLS. That includes broader query helpers, subledger workflows, policy/governance controls, webhook/backup automation, and principal/runtime admin flows.

First run

REVURA_DB_PATH=./books/general.db \
REVURA_ENTITY_ID=entity:demo \
REVURA_MASTER_ROOT=00112233445566778899aabbccddeeff00112233445566778899aabbccddeeff \
revura-native-gateway

That starts the default stdio/local-native surface.

If you need a service boundary, prefer loopback-only HTTP (127.0.0.1) and treat non-loopback/networked HTTP as advanced deployment work.

Binaries

• revura-native-gateway
• revura-native-principal-keygen
• revura-native-validate-identity-policy
• revura-native-validate-principal-policy

Docs

• Native cookbook: https://github.com/rev1ck/revura/blob/main/docs/native-tool-cookbook.md
• Active execution plan: https://github.com/rev1ck/revura/blob/main/docs/execution-plan.md