English | 日本語 | 简体中文 | 繁體中文 | 한국어 | Español | Русский | Français | Deutsch | Português
Future management layer for the AO public set. Start with
ao-operator; this repo is
for teams that already need typed state across many AO runs.
AO Control Plane turns AO Runtime and AO Operator runs into typed, assumption-aware, replayable work records.
This repository is the first MVP slice. It does not replace AO Runtime or AO Operator. It ingests profile runs, writes typed run-state records, checks that the record is internally coherent, and compiles a portable control-plane evidence tar.
The repos are public together, but the advertisement should still name one product: AO Operator.
| Repo | What it is | How to read it |
|---|---|---|
ao-operator |
The product: role contracts, RunSpecs, provider routing, evidence packs, release gates | Start here. Clone this first. |
ao-runtime |
The Rust execution engine: DAG scheduler, policy seam, event log, artifacts, workers, adapters | Read when you want to understand or embed the engine. |
financial-services-profile |
Flagship demo profile: citation-sensitive financial workflows over public/synthetic data | Run after AO Operator to see the regulated-workflow story. |
secure-agent-profile |
Reusable secure coding-agent profile: guarded patching, dependency review, PR evidence | Run after AO Operator to see policy-gated software work. |
ao-control-plane |
Future management layer: typed run state, evidence aggregation, release-train gates | You are here. Read this after the product and profiles. |
AO Control Plane is not required to install, run, or evaluate AO Operator. It is the roadmap layer for typed run state, evidence aggregation, and release-train control once usage proves the need.
The MVP can ingest secure-agent, financial-services, and AO Operator proof runs. For local development, point the commands below at a run directory produced by one of the profile repos.
Use sample run paths from your own checkout; do not commit local absolute paths into public examples.
python -m ao_control_plane.cli init --repo ../secure-agent-profile
python -m ao_control_plane.cli ingest secure-agent \
--repo ../secure-agent-profile \
--run ../secure-agent-profile/runs/<secure-agent-run-id>
python -m ao_control_plane.cli status --state .ao-control/runs/<secure-agent-run-id>/run-state.json
python -m ao_control_plane.cli verify --state .ao-control/runs/<secure-agent-run-id>/run-state.json
python -m ao_control_plane.cli compile-pack --state .ao-control/runs/<secure-agent-run-id>/run-state.jsonFinancial-services proof:
python -m ao_control_plane.cli ingest financial-services \
--repo ../financial-services-profile \
--run ../financial-services-profile/runs/public-sec-citation-proof
python -m ao_control_plane.cli ingest financial-services \
--repo ../financial-services-profile \
--run ../financial-services-profile/runs/<earnings-note-run-id>
python -m ao_control_plane.cli ingest financial-services \
--repo ../financial-services-profile \
--run ../financial-services-profile/runs/<kyc-triage-run-id>When the run contains public-proof.json, AO Control Plane preserves it as
artifact-public-proof-summary, adds step-public-proof-summary, and records
the external-publication boundary from the proof summary as a policy decision.
AO Operator-native proof:
python -m ao_control_plane.cli ingest ao-operator \
--repo ../ao-operator \
--run ../ao-operator/run-artifacts/<ao-operator-run-id>Release-train proof:
python -m ao_control_plane.cli manifest \
--output .ao-control/release-train/current/release-train-manifest.json
python -m ao_control_plane.cli train-gate \
--output .ao-control/release-train/currentRelease tag enforcement:
python -m ao_control_plane.cli train-gate \
--output .ao-control/release-train/current \
--require-target-tags \
--verify-tag-signatures \
--ao-operator-run /path/to/tagged/ao-operator/run \
--secure-agent-run /path/to/tagged/secure-agent/run \
--financial-services-public-proof-run /path/to/financial-services/public-proof-run \
--financial-services-earnings-run /path/to/tagged/financial-services/earnings-run \
--financial-services-kyc-run /path/to/tagged/financial-services/kyc-runThe train gate ingests the AO Operator, secure-agent, financial-services
public-proof, earnings-note, and KYC sample runs; verifies each typed state;
compiles per-run control-plane evidence packs; and then compiles one
ao-release-train-evidence.tar bundle. The --require-target-tags
mode fails until every product member has its configured release tag; adding
--verify-tag-signatures also requires Git to verify those tags locally. The
five --*-run flags let the final gate consume release-blocking fixture outputs
from tagged artifacts instead of the built-in sample run paths.
python -m pytest -q