feat: Publish 2.0 control-plane readiness summary

README.md

@@ -26,6 +26,22 @@ a Jira-like lifecycle that coding agents can operate safely:
 - `ticket finish` merges only when policy allows, then converges issue state.
 - `doctor` detects repo, workflow, label, and closed-issue status drift.
 
+The 2.0 control-plane surface extends that loop with stable contracts for
+long-running agent work:
+
+- `goal status`, `goal next`, `goal plan`, and `goal finish` model a larger
+  objective as an evidence-backed child-ticket graph with explicit human-review
+  stop conditions.
+- Ticket readiness, PR readiness, review packets, finish readiness, finish
+  receipts, and drift audits make completion reviewable without trusting an
+  agent transcript.
+- `workspace status --json` includes `workspace-queues/v1` so operators can see
+  agent-ready, review-needed, finish-ready, blocked, failed-check, and
+  human-decision queues across repos.
+- Adapter-facing command metadata and approval evidence, including
+  `gira-approval-plan/v1`, let durable agent runtimes treat Gira dry-runs as
+  auditable plans before any matching `--apply`.
+
 The core safety model is Terraform-like:
 
 - `--dry-run` previews GitHub or repository mutations.

docs/agent-kernel-adapter-contract.md

@@ -229,10 +229,11 @@ The approval packet should include:
 An approval for one command must not authorize a different command, a different
 repo, a different issue, a different branch, or additional flags.
 
-## Current Gaps
+## Remaining Hardening
 
-The first flow can run with existing commands, but these gaps should be fixed
-before broad adapter use:
+The 2.0 control-plane path can run with stable command capability metadata,
+schema-versioned goal/workspace/readiness surfaces, and shared approval evidence
+for the main Gira apply boundaries. These gaps remain before broad adapter use:
 
 | Gap | Impact | Follow-up |
 | --- | --- | --- |

docs/product-os-roadmap.md

@@ -8,6 +8,39 @@ This roadmap keeps three constraints fixed:
 2. All new Product OS automation is dry-run-first.
 3. Capability/permission checks gate apply behavior before any mutation.
 
+## 2.0 Control-Plane Readiness
+
+As of 2026-05-23, the 2.0 core is implemented as a CLI-first control-plane
+contract rather than a UI or hosted service. The public product framing is:
+
+> A GitHub-native control plane that turns human and AI work into reviewable,
+> auditable, evidence-backed completion.
+
+The stable 2.0 contract areas are:
+
+| Area | Ready surface |
+| --- | --- |
+| State-aware ticket runtime | Current ticket/PR resolution, structured ticket status JSON, action-oriented next steps, role packets, review packets, and lifecycle dry-run/apply commands. |
+| Evidence-based finish and audit | Ticket readiness, PR readiness, finish readiness, finish receipts, telemetry/provenance guidance, and workflow drift audit reports. |
+| Goal mode | `goal-status/v1`, `goal-next/v1`, `goal-plan/v1`, `goal-finish-readiness/v1`, and idempotent human-review handoff receipts over child ticket graphs. |
+| Workspace health | `workspace-queues/v1` in `workspace status --json`, covering agent-ready, review-needed, finish-ready, blocked, failed-check, and human-decision queues. |
+| Agent adapter boundary | Command capability metadata, schema-versioned JSON surfaces, and shared `gira-approval-plan/v1` approval evidence for matching Gira apply boundaries. |
+
+Remaining hardening is intentionally narrower than the 2.0 feature definition:
+
+- Add JSON contracts or explicit adapter-unsupported markings to legacy
+  text-first command families.
+- Add explicit `post_apply_verification` fields to apply reports that still
+  require command-specific follow-up knowledge.
+- Keep Jira transition planning as `jira-transition-plan/v1` read-only evidence
+  unless a future, explicitly approved Jira mutation boundary is designed.
+- Leave hosted dashboards, broad background sync, and UI workflows to the 3.0
+  and hosted-control-plane tracks.
+
+Historical child tickets that predate finish receipts still require maintainer
+judgment in #521. Do not invent backfilled evidence for those issues; use the
+existing goal-finish human-review handoff when closing the planning loop.
+
 ## Current Baseline
 
 Shipped before this roadmap: