@event4u/agent-config 9.28.0 → 9.29.0

package/.claude-plugin/marketplace.json

@@ -6,7 +6,7 @@
   },
   "metadata": {
     "description": "Bootstrap shim for the event4u agent-config suite. This plugin ships ONLY the dispatcher hooks plus one install-pointer skill — ALL content (skills, rules, commands) is installed via `npx -y @event4u/agent-config init`, which writes the file projection AND registers the same hooks in ~/.claude/settings.json. DEPRECATED as a content channel; kept so the marketplace listing stays discoverable and existing installs keep working hooks.",
-    "version": "9.28.0",
+    "version": "9.29.0",
     "keywords": [
       "agent-config",
       "skills",

package/AGENTS.md

@@ -1,6 +1,6 @@
 # event4u/agent-config
 
-> Shared skill / rule / command suite for AI coding tools. No app runtime. AGENTS.md of the **package itself**.
+> Shared skill / rule / command suite for AI coding tools. No app runtime.
 
 ## Source of truth
 
@@ -9,10 +9,10 @@ Edit `src/` only (`skills`, `rules`, `agent-src`). `task sync` + `task generate-
 ## Working on this repo
 
 ```bash
-task sync            # dist/agent-src/, .augment/
-task generate-tools  # .claude/, .cursor/, .clinerules/, .windsurfrules
-task ci              # full pipeline, green before PR
-# maintainer: AGENT_CONFIG_DEV_MODE=1 -> --scope=project (docs/maintainers/dev-mode.md)
+task sync # dist/agent-src/, .augment/
+task generate-tools # .claude/, .cursor/, .clinerules/, .windsurfrules
+task ci # full pipeline, green before PR
+# dev: AGENT_CONFIG_DEV_MODE=1 -> --scope=project (docs/maintainers/dev-mode.md)
 ```
 
 ## Pointers
@@ -28,7 +28,7 @@ task ci              # full pipeline, green before PR
 - **Discovery** — workspaces/packs: [`ADR-013`](docs/decisions/ADR-013-discovery-frontmatter-contract.md) + [`customization`](docs/customization.md).
 - **Root & `agents/` layout** — `src/` = source ([`ADR-050`](docs/decisions/ADR-050-workspace-vs-package-root-boundary.md)); contract: [`agents-layout`](docs/contracts/agents-layout.md).
 - **Security** — [`SECURITY.md`](SECURITY.md); [`docs/threat-model.md`](docs/threat-model.md).
-- **Delegate & review**: subagents take independent-slice work, not serial; no mutating session ends without neutral review — [`delegation-policy`](dist/agent-src/rules/delegation-policy.md).
+- **Delegate & review**: subagents take independent-slice work, not serial; in a team (host-flag gated) the lead coordinates, teammates implement; no mutating session ends without neutral review — [`delegation-policy`](dist/agent-src/rules/delegation-policy.md).
 
 ## Emergency triage — when nothing else is reachable
 

package/CHANGELOG.md

@@ -312,6 +312,72 @@ Entry-shape contract: [`docs/contracts/CHANGELOG-conventions.md`](docs/contracts
 > that forces a new era split (`# Era: 9.28.x`, etc.) — see
 > [`docs/contracts/CHANGELOG-conventions.md § Era splits`](docs/contracts/CHANGELOG-conventions.md).
 
+## [9.29.0](https://github.com/event4u-app/agent-config/compare/9.28.0...9.29.0) (2026-08-09)
+
+### Release highlights
+
+<!-- Curated head: fill before merge, keep it under 10 lines, and leave `_none_` where it is genuinely the answer. The generated log below is unchanged. -->
+- **Behaviour changes:** _auto-derived, rewrite before merge:_ rule/schema diffs, breaking commits or removed public surface in 1a2dd51, 0304406, 1008212.
+- **Default changes + migration:** _auto-derived, rewrite before merge:_ commits naming a default, migration or migrate in 51088bd, 2590391, b7b4cb0.
+- **Security and correctness:** _none_
+- **Honest nulls:** _none_
+- **Known limitations:** _none_
+
+### Features
+
+* **team:** availability replaces the master switch; teams readiness ([a94a91a](https://github.com/event4u-app/agent-config/commit/a94a91a30b916909c391939320ae4fb2b04e0b31))
+* **council:** CLI-first shipped default, quorum, absent members, handoff ([51088bd](https://github.com/event4u-app/agent-config/commit/51088bdaaf3633c4918ea05182c5becf65593cf2))
+* **dispatch:** judgment ladder replaces the activation gate ([1a2dd51](https://github.com/event4u-app/agent-config/commit/1a2dd5177b02b9ca752588a022d8a72c6afe51f5))
+* **settings:** activation gates die, one audited incident switch survives ([0304406](https://github.com/event4u-app/agent-config/commit/0304406a536e29d66f44ce19f5f0187383b8af4e))
+* **ci:** lint_never_silent - no shipped guidance may direct a silent re-run ([c7b3d83](https://github.com/event4u-app/agent-config/commit/c7b3d8382ac34a992727c1a042cc65298e56adaa))
+* **self-repair:** fork-aware egress ladder, issue form, detector corpus gate ([5b50807](https://github.com/event4u-app/agent-config/commit/5b5080710e9800bbb4e1eec5773d9c5ff79de866))
+* **self-repair:** fork-aware egress, detector corpus gate, never-silent lint ([4e046aa](https://github.com/event4u-app/agent-config/commit/4e046aa5d992cbefa31717430bd05a9e37798f4f))
+* **rules:** emit Claude Code rules with the host-native paths key ([872c267](https://github.com/event4u-app/agent-config/commit/872c267589995668bfbfffdc2c0128b9d89532dd))
+* **forensics:** read-only forensic analyzers as a default-off pack (Phase 3) ([b7b4cb0](https://github.com/event4u-app/agent-config/commit/b7b4cb05e2247737786e96ff52cf4d739dff68c4))
+* **bench:** critic-protocol A/B result — no promotion, published both directions (Phase 2) ([bc0054b](https://github.com/event4u-app/agent-config/commit/bc0054b8406dc6fe280fcca7b10104547e2380b1))
+* **council:** critic_protocol config + pre-registered load_bearing A/B harness (Phase 2, pre-run) ([e75ae51](https://github.com/event4u-app/agent-config/commit/e75ae514b765205a2f86c5ffaae40a5c81523d87))
+* **hosts:** add worker_respawn to the host-capability manifest ([a30102d](https://github.com/event4u-app/agent-config/commit/a30102d0f30d197ced06248c5712722006032447))
+* **subagents:** emit the capsule at a budget watermark, in shadow ([572d280](https://github.com/event4u-app/agent-config/commit/572d28066be1871f43242056af175ddbd023002c))
+* **subagents:** add the CHECKPOINT capsule as the worker handoff shape ([8340f5e](https://github.com/event4u-app/agent-config/commit/8340f5e11219c296030084432e9f403818552df9))
+* **premortem:** four-part failure register + optional roadmap Pre-mortem section (Phase 1) ([da95e9e](https://github.com/event4u-app/agent-config/commit/da95e9e4880394ab06ea0aaf818e893c695d4657))
+
+### Bug Fixes
+
+* **build:** rebuild the committed install bundle for the settings teardown ([6d9e667](https://github.com/event4u-app/agent-config/commit/6d9e6672bee69178d0ce8d86bd574e248abca54b))
+* **spike:** escape the shape-key separator instead of a raw NUL byte ([1dd9a28](https://github.com/event4u-app/agent-config/commit/1dd9a281d5fc38ce228a04035a43a35b295114e5))
+* **self-repair:** resolve the four R2 completion-review findings ([3474bf7](https://github.com/event4u-app/agent-config/commit/3474bf797321a99e5a7dd5a484a70a8a8748add6))
+* **roadmap:** risk-register marker, allowed risk types, resolvable anchors ([f06f6ed](https://github.com/event4u-app/agent-config/commit/f06f6edb1f9f3b08c4763c542e789272f4dbd091))
+* **scorer:** repoint skill-selection scoring at the live skills tree ([589b2c3](https://github.com/event4u-app/agent-config/commit/589b2c3e11eaf64a32e7bc7269807b3a9af648a2))
+* **ci:** unlink the ADR-217 body link that can never resolve in the dist projection ([1008212](https://github.com/event4u-app/agent-config/commit/100821272c9254d4aae70876c9bfa65af67d1d4d))
+* **forensics:** review findings — schema-valid finding ids, fail-before-write, hermetic smoke test, honest config scope ([0509647](https://github.com/event4u-app/agent-config/commit/050964793ec76e0100199a59fefd5b8cc4ad59e4))
+* **release:** tolerate the concurrent tag-push race in step 8 ([049e8b2](https://github.com/event4u-app/agent-config/commit/049e8b24296a97f946c4597418cfd79457018e3c))
+* **hooks:** self-heal a stale hook bundle in preflight instead of blocking the push ([9256633](https://github.com/event4u-app/agent-config/commit/9256633996be6d86e09df8c288837a8970d456e7))
+
+### Documentation
+
+* **council:** describe the runtime responses artefact without the pruned path literal ([04e8a0f](https://github.com/event4u-app/agent-config/commit/04e8a0feecb2fc19326f9abf894cf49f38424c6b))
+* **roadmap:** always-on orchestration - verified claims, council cut, honest blockers ([d202556](https://github.com/event4u-app/agent-config/commit/d202556efa32376c4fc5fedba7fb95414cb4f4b4))
+* always-on doctrine reaches the always-loaded surfaces ([9eead89](https://github.com/event4u-app/agent-config/commit/9eead8923e604f584541afbdac6d1feac3b02e5c))
+* **review:** re-derive the context-manifest scope_hash for the re-bound round (contract 5 header-manifest agreement) ([2ebdb87](https://github.com/event4u-app/agent-config/commit/2ebdb87f0286a167d24cc26b3a658306274748c6))
+* **review:** add fix-commit refs to the four resolved R2 findings ([9adbf83](https://github.com/event4u-app/agent-config/commit/9adbf83bcb530db4d702d4ff9425f685124b8a2f))
+* **review:** re-bind R2 findings to the post-fix review scope (contract 2.1 in-place re-bind) ([752c29c](https://github.com/event4u-app/agent-config/commit/752c29cf11a66ded030bff60aeac45763c36c2ed))
+* **review:** R2 completion-review findings for feat/road-to-rule-delivery-integrity (1 medium, 3 low, all open) ([0267568](https://github.com/event4u-app/agent-config/commit/02675686ef102552a80ce35861e1c5c094fb3f31))
+* **roadmap:** close rule-delivery-integrity, migrate the measurement out ([2590391](https://github.com/event4u-app/agent-config/commit/25903915772394a776b069991e3d939bcf3f0e70))
+* **adr:** ADR-220 skill invocation attestation — specified, check deferred (Phase 4) ([c8f3dbb](https://github.com/event4u-app/agent-config/commit/c8f3dbb31762a7df7ef3abf7ef70f1e17741afa4))
+* **roadmap:** land phases 0 and 1 of worker-generation recycling ([06c7347](https://github.com/event4u-app/agent-config/commit/06c7347b49809a81b7ee169f42e3c742e5211812))
+* **claims:** pre-register the capsule trigger-arm comparison ([2acfc56](https://github.com/event4u-app/agent-config/commit/2acfc56982d257331243094d2a580ca4f5926c6c))
+
+### Chores
+
+* **roadmap:** close Phase 4 of road-to-rule-delivery-integrity ([5db59bc](https://github.com/event4u-app/agent-config/commit/5db59bc4410488eb22717c8d0606b8eca10bcfa4))
+* **roadmap:** park round-6 and worker-generation-recycling in later/ ([e5ebedc](https://github.com/event4u-app/agent-config/commit/e5ebedcd9ba0b3ddaba10938af74fcab9d79f03a))
+* **generated:** regenerate proof.md + CAPABILITIES.yaml for the 289th skill post-merge ([d0f348d](https://github.com/event4u-app/agent-config/commit/d0f348d0b75b507d10199941185969a633563f63))
+* **roadmap:** close Phase 5 + acceptance criteria — roadmap fully closed ([c993660](https://github.com/event4u-app/agent-config/commit/c9936607fe77e06f1b510a3ee3ccf8390f49a8d6))
+* **generated:** rebuild the install bundle for the settings-schema change ([5b1260d](https://github.com/event4u-app/agent-config/commit/5b1260de7b889a16e94a5e5d5bc6f22f23dc558e))
+* **generated:** regenerate proof.md for the new ledger entry ([76ab0a9](https://github.com/event4u-app/agent-config/commit/76ab0a9e1ccdb09e436f562c1cf8416989d921ef))
+
+Tests: 12438 (+297 since 9.28.0)
+
 ## [9.28.0](https://github.com/event4u-app/agent-config/compare/9.27.0...9.28.0) (2026-08-09)
 
 ### Release highlights

package/README.md

@@ -4,7 +4,7 @@
 
 [![Smoke](https://github.com/event4u-app/agent-config/actions/workflows/smoke.yml/badge.svg)](https://github.com/event4u-app/agent-config/actions/workflows/smoke.yml) [![Public install smoke (3 OS × 2 Node)](https://github.com/event4u-app/agent-config/actions/workflows/smoke-public-install.yml/badge.svg)](https://github.com/event4u-app/agent-config/actions/workflows/smoke-public-install.yml) [![npm](https://img.shields.io/npm/v/@event4u/agent-config?style=flat-square&label=npm&color=orange)](https://www.npmjs.com/package/@event4u/agent-config) [![agent-config MCP server](https://glama.ai/mcp/servers/event4u-app/agent-config/badges/score.svg)](https://glama.ai/mcp/servers/event4u-app/agent-config) [![MCP Toplist](https://mcptoplist.com/badge/glama%2Fevent4u-app%2Fagent-config.svg)](https://mcptoplist.com/server/glama%2Fevent4u-app%2Fagent-config)
 
-[![Skills](https://img.shields.io/badge/Skills-288-orange?style=flat-square)](dist/agent-src/skills/) [![Rules](https://img.shields.io/badge/Rules-116-orange?style=flat-square)](dist/agent-src/rules/) [![Commands](https://img.shields.io/badge/Commands-196-orange?style=flat-square)](dist/agent-src/commands/) [![Guidelines](https://img.shields.io/badge/Guidelines-104-orange?style=flat-square)](docs/guidelines/) [![Personas](https://img.shields.io/badge/Personas-29-orange?style=flat-square)](dist/agent-src/personas/) [![Advisors](https://img.shields.io/badge/Advisors-5-orange?style=flat-square)](dist/agent-src/personas/advisors/)
+[![Skills](https://img.shields.io/badge/Skills-289-orange?style=flat-square)](dist/agent-src/skills/) [![Rules](https://img.shields.io/badge/Rules-116-orange?style=flat-square)](dist/agent-src/rules/) [![Commands](https://img.shields.io/badge/Commands-196-orange?style=flat-square)](dist/agent-src/commands/) [![Guidelines](https://img.shields.io/badge/Guidelines-104-orange?style=flat-square)](docs/guidelines/) [![Personas](https://img.shields.io/badge/Personas-29-orange?style=flat-square)](dist/agent-src/personas/) [![Advisors](https://img.shields.io/badge/Advisors-5-orange?style=flat-square)](dist/agent-src/personas/advisors/)
 
 **Try it in 30 seconds** — drop one read-only subagent into any repo and watch it gate "done": `@production-validator check this branch is actually done`. No wizard, no lock-in, nothing else installed — the [30-second wedge ↓](#quickstart) is the whole first step. Start at the proof, not the catalog: **[event4u-app.github.io/agent-config/proof/](https://event4u-app.github.io/agent-config/proof/)**.
 

package/agents/templates/.ai-council.yml.example

@@ -65,21 +65,30 @@ defaults:
   #            reason. `manual` is NEVER part of the auto chain (it would
   #            always "succeed" and silently turn every call into a
   #            copy-paste request) — it stays an explicit opt-in.
-  #            OPT-IN: `auto` is not the shipped default. Switching to it
-  #            moves spend from per-token dollars onto your subscription
-  #            quota, so it is a decision you take, not one you inherit.
-  #            Set `cli_call_budget` below before you flip it.
+  #            THE SHIPPED DEFAULT (road-to-always-on-orchestration
+  #            Phase 3.1 — CLI-first is the owner-set transport doctrine):
+  #            ride the vendor CLI under your subscription first, fall to
+  #            the metered API only where no CLI resolves. Pin `manual` /
+  #            `api` / `cli` here (or per-member below) to override it.
+  #            `cli_call_budget` below ships populated with a generous
+  #            per-provider guard so this default never runs your
+  #            subscription quota uncapped.
   #
   # Precedence: invocation flag > per-member mode > defaults.mode >
   # built-in. Two DIFFERENT defaults sit at the bottom, and they are not
   # the same thing:
-  #   · omit `defaults.mode` entirely and the loader fills "api" — that is
-  #     what every real config file observes.
+  #   · omit `defaults.mode` entirely and the loader fills "auto" (was
+  #     "api" before the Phase 3.1 flip) — that is what every real config
+  #     file observes.
   #   · the resolver's built-in fallback, reached only when no layer at all
   #     supplies a mode (a settings dict handed straight to the resolver,
   #     no config file involved), is "manual" — free, no key. A caller who
-  #     named no transport has not asked to spend money.
-  mode: api
+  #     named no transport has not asked to spend money. This did NOT flip
+  #     alongside the loader default: `auto` still resolves to a paying
+  #     rung when no CLI is usable, so flipping this too would let a
+  #     caller that bypassed the config file spend money on a preference
+  #     it never stated.
+  mode: auto
 
   # Default number of debate rounds per /council invocation.
   #
@@ -179,26 +188,34 @@ debate:
     threshold_usd: 1.00
     show_per_member: true
 
-# Per-day call-count guard for `mode: cli` members. The standard
-# `cost_budget` block skips CLI calls because they are billable=False —
-# but provider subscriptions still carry their own quotas (Claude Pro
-# 5h windows, ChatGPT Plus message caps, Gemini free-tier per-day
-# limits). Opt into a per-provider cap; default unset = unlimited from
-# this loader's perspective.
+# Per-day call-count guard for `mode: cli` / `mode: auto` members. The
+# standard `cost_budget` block skips CLI calls because they are
+# billable=False — but provider subscriptions still carry their own quotas
+# (Claude Pro 5h windows, ChatGPT Plus message caps, Gemini free-tier
+# per-day limits). The loader SHIPS EVERY known provider populated at
+# 50/day by default (road-to-always-on-orchestration Phase 3.4) — a config
+# that omits this block entirely, or omits one provider from the map
+# below, still gets that floor for the missing entries. It is a GUARD
+# against an always-on `auto` default silently exhausting a subscription's
+# plan quota, not a brake — override a provider explicitly to tighten it
+# (worked example below) or loosen it further.
 #
 # Counter state persists at ~/.event4u/agent-config/cli-calls.json
 # with daily UTC reset. Wired by Phase 1 of the CLI-transport roadmap.
 #
-# SHIPPED POPULATED, deliberately: `mode: auto` prefers the cli rung, so
-# shipping that path with no quota guard would miss the point. The caps
-# below are sized from the worst case one invocation can cost — a deep
-# debate is at most `debate_max_rounds` (4) calls per member — so 24
-# permits roughly six full deep debates per provider per day before the
-# guard fires. That is well inside a Pro-tier window while still bounding
-# a runaway loop, and failures count too (a wedged CLI cannot burn the
-# whole budget retrying). Gemini's free tier is the tightest, hence the
-# lower cap. Raise them if you consult heavily; `council:quota` reports
-# usage and `council:quota --reset` clears the day.
+# The three entries below TIGHTEN the 50/day floor, deliberately: `mode:
+# auto` prefers the cli rung, so a config that consults heavily wants a
+# number sized to its own worst case, not the generic guard. Sized from
+# the worst case one invocation can cost — a deep debate is at most
+# `debate_max_rounds` (4) calls per member — so 24 permits roughly six
+# full deep debates per provider per day before the guard fires. That is
+# well inside a Pro-tier window while still bounding a runaway loop, and
+# failures count too (a wedged CLI cannot burn the whole budget retrying).
+# Gemini's free tier is the tightest, hence the lower cap. xai and
+# perplexity are left unlisted here and so keep the loader's 50/day floor —
+# raise or lower any of the five if you consult heavily or want a tighter
+# rein; `council:quota` reports usage and `council:quota --reset` clears
+# the day.
 cli_call_budget:
   max_calls_per_day:
     anthropic: 24
@@ -207,6 +224,20 @@ cli_call_budget:
   # Warn once usage crosses this fraction of a cap (0..1).
   warn_at: 0.8
 
+# Quorum for a council pass to CONCLUDE (road-to-always-on-orchestration
+# Phase 3.3). Default "majority" resolves to ceil(n / 2) enabled members —
+# a SIMPLE majority, deliberately not "more than half": at n=2 that is
+# 1-of-2, because requiring 2-of-2 turns any single absent member into a
+# deadlocked release gate. Below quorum the pass is `inconclusive` and, at
+# a release gate, HOLDS for a human — never silently downgraded to
+# advisory. Set an explicit integer to require a fixed count instead:
+#
+#   quorum: 2
+#
+# See docs/contracts/ai-council-config.md § Quorum for the full resolver
+# contract and the manifest.json / rendered-report fields it feeds.
+quorum: majority
+
 # Per-member configuration — you decide what may spend; detection tells
 # you what is already here.
 #

package/dist/agent-src/commands/cost/report.md

@@ -159,8 +159,8 @@ If no orchestration lines exist yet (telemetry capture not yet running or
 no delegated dispatches this month), surface:
 
 > No orchestration telemetry yet — subagent dispatches begin writing to
-> `agents/runtime/state/audit/YYYY-MM.jsonl` once `subagents.enabled: true`
-> and a dispatch runs.
+> `agents/runtime/state/audit/YYYY-MM.jsonl` once a dispatch runs (always-on
+> orchestration: the layer has no per-run on/off setting any more).
 
 Skip this step entirely if the user did not ask about orchestration telemetry
 and there is no `agents/runtime/state/audit/` directory.

package/dist/agent-src/commands/team/adversarial.md

@@ -29,11 +29,11 @@ attacker — a second strong model with repo access instead of the host model
 critiquing itself. Use it as the **escalation rung**: single-model
 self-review found nothing, or the stakes are high.
 
-### 1. Gate — `ai_team.enabled`
+### 1. Gate — `/team` availability
 
-Read `ai_team.enabled` from `.agent-settings.yml`. Missing or `false` →
-print the enable pointer from `/team` (master) § "Default-off gate" and
-**STOP**.
+Run the availability check from `/team` (master) § "Availability gate":
+codex CLI + auth, and `emergency.orchestration_halt`. Either check fails →
+print the matching block and **STOP**.
 
 ### 2. Gate — plugin presence (fail closed)
 
@@ -73,12 +73,13 @@ on quota exhaustion surface the transport's refusal, do not retry.
 ## Output format
 
 - The plugin's adversarial findings, verbatim.
-- Gate failures print exactly one block (enable pointer or fail-closed
-  block) and stop.
+- Gate failures print exactly one block (availability/halt block or
+  fail-closed block) and stop.
 
 ## Do NOT
 
-- Do NOT run when `ai_team.enabled` is false — enable pointer, stop.
+- Do NOT run when `/team` is unavailable (codex CLI/auth missing) or
+  halted — matching block, stop.
 - Do NOT reimplement the adversarial pass inline when the plugin is absent —
   fail closed with the doctor pointer.
 - Do NOT skip the free first rung reflexively — the single-model

package/dist/agent-src/commands/team/delegate.md

@@ -27,16 +27,16 @@ repo. On Claude Code hosts it delegates to the official plugin's
 `/codex:rescue`. This is the **only wrapper that delegates write access**,
 so it carries a second opt-in on top of the family gate.
 
-### 1. Gate — `ai_team.enabled`
+### 1. Gate — `/team` availability
 
-Read `ai_team.enabled` from `.agent-settings.yml`. Missing or `false` →
-print the enable pointer from `/team` (master) § "Default-off gate" and
-**STOP**.
+Run the availability check from `/team` (master) § "Availability gate":
+codex CLI + auth, and `emergency.orchestration_halt`. Either check fails →
+print the matching block and **STOP**.
 
 ### 2. Gate — `ai_team.allow_delegate` (second opt-in)
 
 Read `ai_team.allow_delegate`. Missing or `false` (the shipped default) →
-refuse with the enable pointer and **STOP**:
+refuse with the opt-in pointer and **STOP**:
 
 > `/team delegate` is disabled (`ai_team.allow_delegate: false`, the
 > default). It is the only team-mode wrapper that hands **write access** to
@@ -44,15 +44,15 @@ refuse with the enable pointer and **STOP**:
 >
 > ```yaml
 > ai_team:
->   enabled: true
 >   allow_delegate: true
 > ```
 >
 > Key reference: `docs/contracts/ai-team-config.md`.
 
 Deterministic mirror of gates 1+2: `npx tsx src/scripts/ai_team/team_dispatch.ts
---delegate-gate` exits non-zero (with the pointer on stderr) unless BOTH flags
-are true — run it instead of hand-parsing the YAML when a shell is available.
+--delegate-gate` exits non-zero (with the pointer on stderr) unless `/team`
+is available AND `allow_delegate` is true — run it instead of hand-parsing
+the YAML when a shell is available.
 
 ### 3. Gate — plugin presence (fail closed)
 
@@ -91,14 +91,15 @@ opens a PR on top of them.
 ## Output format
 
 - The plugin's rescue/delegation result, verbatim.
-- Gate failures print exactly one block (enable pointer, allow_delegate
-  refusal, or fail-closed block) and stop.
+- Gate failures print exactly one block (availability/halt block,
+  allow_delegate refusal, or fail-closed block) and stop.
 
 ## Do NOT
 
-- Do NOT run when `ai_team.enabled` is false — enable pointer, stop.
+- Do NOT run when `/team` is unavailable (codex CLI/auth missing) or
+  halted — matching block, stop.
 - Do NOT run when `ai_team.allow_delegate` is false — refusal block, stop.
-  Never treat `enabled: true` alone as delegate authorization.
+  Never treat availability alone as delegate authorization.
 - Do NOT reimplement the worker inline when the plugin is absent — fail
   closed with the doctor pointer.
 - Do NOT commit, push, or open a PR over the worker's edits — review is the

package/dist/agent-src/commands/team/review.md

@@ -6,7 +6,7 @@ argument-hint: "[--background]"
 pack: meta
 tier: 2
 visibility: internal
-description: Thin wrapper — cross-model review of the current diff via the official plugin (/codex:review). Gated on ai_team.enabled; fails closed when the plugin is absent.
+description: Thin wrapper — cross-model review of the current diff via the official plugin (/codex:review). Gated on /team availability (codex CLI + auth); fails closed when the plugin is absent.
 cluster: team
 sub: review
 suggestion:
@@ -26,11 +26,11 @@ Thin wrapper: cross-model review of the current change. On Claude Code hosts
 it delegates verbatim to the official plugin's `/codex:review` — this wrapper
 owns only the governance gates around it.
 
-### 1. Gate — `ai_team.enabled`
+### 1. Gate — `/team` availability
 
-Read `ai_team.enabled` from `.agent-settings.yml`. Missing or `false` →
-print the enable pointer from `/team` (master) § "Default-off gate" and
-**STOP**. No plugin probe, no partial run.
+Run the availability check from `/team` (master) § "Availability gate":
+codex CLI + auth, and `emergency.orchestration_halt`. Either check fails →
+print the matching block and **STOP**. No plugin probe, no partial run.
 
 ### 2. Gate — plugin presence (fail closed)
 
@@ -65,12 +65,13 @@ refusal, do not retry.
 ## Output format
 
 - The plugin's review, verbatim.
-- Nothing else on success; gate failures print exactly one block (enable
-  pointer or fail-closed block) and stop.
+- Nothing else on success; gate failures print exactly one block
+  (availability/halt block or fail-closed block) and stop.
 
 ## Do NOT
 
-- Do NOT run when `ai_team.enabled` is false — enable pointer, stop.
+- Do NOT run when `/team` is unavailable (codex CLI/auth missing) or
+  halted — matching block, stop.
 - Do NOT reimplement the review inline when the plugin is absent — fail
   closed with the doctor pointer.
 - Do NOT edit files, commit, push, or open a PR — this wrapper is read-only.

package/dist/agent-src/commands/team/status.md

@@ -5,7 +5,7 @@ disable-model-invocation: true
 pack: meta
 tier: 2
 visibility: internal
-description: Thin wrapper — plugin job status via /codex:status plus a quota block (shared openai counter vs team + council ceilings). Gated on ai_team.enabled; fails closed without the plugin.
+description: Thin wrapper — plugin job status via /codex:status plus a quota block (shared openai counter vs team + council ceilings). Gated on codex CLI/auth availability; fails closed without the plugin.
 cluster: team
 sub: status
 suggestion:
@@ -25,11 +25,11 @@ Thin wrapper: team-mode status. On Claude Code hosts it delegates to the
 official plugin's `/codex:status` and appends **our** quota block —
 the one piece of state the plugin does not know about.
 
-### 1. Gate — `ai_team.enabled`
+### 1. Gate — `/team` availability
 
-Read `ai_team.enabled` from `.agent-settings.yml`. Missing or `false` →
-print the enable pointer from `/team` (master) § "Default-off gate" and
-**STOP**.
+Run the availability check from `/team` (master) § "Availability gate":
+codex CLI + auth, and `emergency.orchestration_halt`. Either check fails →
+print the matching block and **STOP**.
 
 ### 2. Gate — plugin presence (fail closed)
 
@@ -102,12 +102,13 @@ compatibility).
   ceilings + per-path OPEN/BLOCKED verdict).
 - Any `team.gate: BLOCK n/N` ledger lines, verbatim, after the quota
   block — only when present.
-- Gate failures print exactly one block (enable pointer or fail-closed
-  block) and stop.
+- Gate failures print exactly one block (availability/halt block or
+  fail-closed block) and stop.
 
 ## Do NOT
 
-- Do NOT run when `ai_team.enabled` is false — enable pointer, stop.
+- Do NOT run when `/team` is unavailable (codex CLI/auth missing) or
+  halted — matching block, stop.
 - Do NOT fabricate a count when the counter file is unreadable — print
   `Quota: unavailable (<reason>)` instead of a guessed number.
 - Do NOT write to `cli-calls.json` — the transport owns the counter; this

package/dist/agent-src/commands/team.md

@@ -15,7 +15,7 @@ type: orchestrator
 suggestion:
   eligible: true
   trigger_description: "second model, GPT drüberschauen lassen, cross-model review, review gate, let another strong model check the diff, team review"
-  trigger_context: "ai_team.enabled is true AND the user wants a second strong model to review the actual working-tree diff with repo access (depth) — not a neutral artefact opinion (that is /council)"
+  trigger_context: "/team is available (codex CLI installed + authenticated, not halted) AND the user wants a second strong model to review the actual working-tree diff with repo access (depth) — not a neutral artefact opinion (that is /council)"
 workspaces:
   - agent-config-maintainer
 packs:
@@ -55,25 +55,33 @@ pre-registered benchmark (§ No-claims note).
 | Model choice | Mid-tier default (cost decision) | Strongest available — `ai_team.model: 'auto'` defers to the CLI default |
 | Write access | Never (text only; one opt-in PR comment under `pr`) | Read-only by default; `/team delegate` is the single write path, behind a second opt-in |
 
-## Default-off gate — check FIRST, before anything else
+## Availability gate — check FIRST, before anything else
 
-Read `ai_team.enabled` from `.agent-settings.yml`. Missing key or `false`
-(the shipped default) → print the enable pointer and **STOP**. Do not route,
-do not suggest a sub-command, do not probe for the plugin:
+`/team` carries no on/off setting — `ai_team.enabled` was REMOVED
+(road-to-always-on-orchestration Phase 1, Step 1.3). Availability is a FACT
+resolved from the machine: the codex CLI must resolve on `$PATH`, AND some
+auth for it must be detectable (subscription login, API key, key file, or
+env key) — the same check `agent-config doctor --check team` runs. Missing
+either → print the degrade line and **STOP**. Do not route, do not suggest
+a sub-command, do not probe for the plugin:
 
-> Team mode is disabled (`ai_team.enabled: false`, the default).
-> Enable it in `.agent-settings.yml`:
->
-> ```yaml
-> ai_team:
->   enabled: true
-> ```
->
-> Key reference: `docs/contracts/ai-team-config.md`. Then run
-> `agent-config doctor --check team` to verify plugin + auth setup.
+> codex CLI not available — `/team` needs the codex CLI installed and
+> authenticated. Install: `npm install -g @openai/codex`, then run
+> `codex login`. Verify with `agent-config doctor --check team`.
+
+Also check the one incident switch: `emergency.orchestration_halt: true`
+in `.agent-settings.yml` halts `/team` along with the rest of the always-on
+orchestration stack (subagents, council, team). Halted → print:
+
+> `/team` is halted (`emergency.orchestration_halt: true`) — the always-on
+> orchestration stack is paused for incident response. Resume: set
+> `orchestration_halt: false` with a non-empty
+> `orchestration_halt_justification` in `.agent-settings.yml`.
 
 Every sub-command repeats this check independently — the gate holds even when
-a wrapper is invoked directly.
+a wrapper is invoked directly. A leftover `ai_team.enabled` key from an
+older install is ignored (one deprecation-warning line on load) — it no
+longer gates anything.
 
 ## Sub-commands
 
@@ -89,7 +97,8 @@ rather than routing to an in-suite skill — the second column names that target
 
 ## Dispatch
 
-1. Run the default-off gate above. Disabled → enable pointer, stop.
+1. Run the availability gate above. Unavailable or halted → the matching
+   one-line block, stop.
 2. Parse the user's argument: `/team <sub-command> [args]`.
 3. Look up the sub-command in the table above; load the routed file and
    follow its `## Instructions` verbatim with the remaining args.
@@ -106,7 +115,7 @@ When the official plugin is absent on a Claude Code host, every wrapper
 fails **closed** with the doctor remediation pointer — never a silent no-op,
 never an inline reimplementation of the plugin:
 
-> ❌ Team mode is enabled, but the official plugin is not installed on this
+> ❌ `/team` is available, but the official plugin is not installed on this
 > Claude Code host. `/team <sub>` fails closed — it never reimplements the
 > plugin inline and never silently no-ops.
 >
@@ -156,4 +165,5 @@ measured improvement.
 - `adversarial-review` skill — the free single-model first rung below
   `/team adversarial`.
 - `docs/contracts/ai-team-config.md` — the `ai_team` settings block
-  (`enabled`, `model`, `allow_delegate`, `max_calls_per_day`).
+  (`model`, `allow_delegate`, `max_calls_per_day`) and the availability
+  contract (codex CLI + auth, `emergency.orchestration_halt`).