agents-city 0.3.0-beta.21

package/README.md

@@ -0,0 +1,1999 @@
+# Agents City
+
+[Español](README.es.md) · [English](README.md)
+
+**Run several autonomous agent cities on one machine, and connect only the
+cities that should talk.**
+
+Agents City is a local-first, multi-model orchestrator for repository work. Each
+city has its own identity, domain, chair seat, goal, repo support agents,
+editable knowledge, live-recognised skills, and explicit roads to other cities.
+It does not turn every agent into a group chat: the seat chairs the process,
+selects specialists, and controls the floor.
+
+This is the complete guide. If you only want to try it, go to
+[Quick start](#quick-start).
+
+## Contents
+
+- [Mental model](#mental-model)
+- [Quick start](#quick-start)
+- [Requirements and installation](#requirements-and-installation)
+- [First run, step by step](#first-run-step-by-step)
+- [Working inside tmux](#working-inside-tmux)
+- [Runtimes and transports](#runtimes-and-transports)
+- [Domains, roles, and knowledge](#domains-roles-and-knowledge)
+- [Complete command reference](#complete-command-reference)
+- [Committee: complete workflow](#committee-complete-workflow)
+- [Claude `/city:` commands](#claude-city-commands)
+- [Use-case cookbook](#use-case-cookbook)
+- [Files and environment variables](#files-and-environment-variables)
+- [Security and trust boundaries](#security-and-trust-boundaries)
+- [Troubleshooting](#troubleshooting)
+- [Development and testing](#development-and-testing)
+
+## Mental model
+
+A city is not an account, a remote person, or a free-form collection of bots. It
+is one autonomous work domain owned by a local person:
+
+```text
+local user
+├── home city
+│   ├── stable identity: owner/home
+│   ├── domain + seat role + goal
+│   ├── seat: chair and only public boundary
+│   ├── agent A: workspace + mounts → a git repo (kind: code)
+│   ├── agent B: workspace + mounts → a folder of documents (kind: knowledge)
+│   ├── agent C: workspace + mounts → several repos and a worktree
+│   ├── editable domain/role knowledge
+│   ├── skills that already live inside the mounted work
+│   └── explicit roads to other seats
+├── product city
+└── client-a city
+```
+
+**Agents come first.** An agent is the unit, and a repo is just one thing it can
+mount. Each agent has a **workspace folder** with a `mounts/` dir of symlinks to
+wherever the real work lives — a git repo, a linked worktree, or a plain folder
+of documents — so a person whose work is knowledge in documents, with no git at
+all, is a first-class agent. "One repo is one agent" is simply the special case
+of an agent whose single mount is that repo, so **existing repo-only cities keep
+working unchanged**. Full model: [docs/agents-first.md](docs/agents-first.md).
+
+The important boundaries are:
+
+- **User:** may own several local cities.
+- **City:** has its own identity, domain, goal, configuration, and state.
+- **Seat:** chairs the committee and is the only actor that may cross roads.
+- **Agent:** the member unit. It owns a workspace folder and works over its
+  mounts, contributes evidence, and always has member authority — whether its
+  speciality is `dev`, `seo` or `cfo`, and whether its **kind** is `code`,
+  `knowledge` or `coordinator`.
+- **Mount:** a symlink inside an agent's workspace to real work on disk (a repo,
+  a worktree, a document folder). An agent may have several, or none.
+- **Role:** professional perspective and responsibility; it does not grant bus
+  permissions.
+- **Skill:** capability installed by the user or repo. Recognition is live and
+  read-only; the one deliberate write is the Hall installing a skill zip the
+  owner explicitly uploads, into that agent's own home — never on its own,
+  never anywhere else. Skills are the Claude runtime's format; other engines
+  ignore them.
+- **Road:** allowlist between two seats. It grants reachability, not authority.
+- **Committee:** bounded process for isolated positions, synthesis, floor,
+  decision, and verification. It is not lateral conversation between all agents.
+
+## Quick start
+
+### Right now: the real npm experience without publishing
+
+Packing first matters: this tests the exact file list a person would receive
+from npm, not the whole checkout.
+
+```bash
+cd /path/to/agents-city
+npm pack
+npm install -g ./agents-city-0.3.0-beta.21.tgz
+agents-city --version
+agents-city seat
+```
+
+You do not need to run `npm publish`. The installation is global only for your
+currently active Node installation.
+
+### Registry installation once available
+
+Check availability and dist-tags first:
+
+```bash
+npm view agents-city dist-tags --json
+```
+
+If it returns `E404`, use the local tarball in the previous section. Once the
+registry shows a `beta` dist-tag, install it with:
+
+```bash
+npm install -g agents-city@beta
+agents-city --version
+agents-city seat
+```
+
+Use `@beta` while this is a prerelease. Bare `npm install -g agents-city` should
+be reserved for the future stable release carrying the `latest` tag.
+
+### Open the Hall instead of the terminal wizard
+
+```bash
+agents-city
+# equivalent to:
+agents-city hall
+```
+
+The Hall listens on `127.0.0.1`, chooses a free port, and opens the browser. You
+can create or select a city and edit its configuration there. The Hall and CLI
+use the same underlying modules.
+
+## Requirements and installation
+
+### Base requirements
+
+| Requirement | Used for |
+|---|---|
+| Node.js 22 or later | npm package, WebSocket bus, and frontends |
+| npm | installation and packaging |
+| Python 3 | Hall, onboarding, cities, maps, and utilities |
+| bash | sessions and launchers |
+| tmux | one window per seat/repo; `seat` tries to install it when missing |
+| macOS or Linux | natively supported platforms |
+| WSL | required on Windows because native Windows has no bash/tmux |
+
+Each runtime also needs its own installed and authenticated CLI. Agents City
+does not bundle or replace Claude, Codex, OpenCode, or Kimi accounts.
+
+```bash
+command -v claude
+command -v codex
+command -v opencode
+command -v kimi
+```
+
+You do not need all of them. An all-Claude, all-Codex, or mixed city is valid.
+
+### GitHub is optional
+
+Selecting local repos requires no account. If you choose GitHub during
+onboarding, Agents City uses the separate `gh` CLI:
+
+1. detects it;
+2. tries the system package manager if it is absent;
+3. runs `gh auth login --web` when unauthenticated;
+4. shows the device code if a browser cannot open;
+5. offers to clone selected repos that are not on disk.
+
+`gh` is not bundled inside the Agents City npm package.
+
+### Update a local installation
+
+```bash
+cd /path/to/agents-city
+npm pack
+npm install -g ./agents-city-0.3.0-beta.21.tgz
+agents-city --version
+```
+
+Already-running sessions retain the code loaded in memory. To apply the new
+version to one city:
+
+```bash
+agents-city exit home --dry-run
+agents-city exit home
+agents-city seat --city home
+```
+
+Save active work first: `exit` closes every window in that city.
+
+### Uninstall
+
+```bash
+npm uninstall -g agents-city
+```
+
+This removes the installed program. It does **not** remove `~/.agents-city`, your
+cities, backups, or repositories. Use `agents-city reset` only when you intend to
+restart one specific city.
+
+## First run, step by step
+
+`agents-city seat` creates `home` when no city exists and asks for five decisions.
+
+### 1. Work domain
+
+The domain determines vocabulary, evidence criteria, and suggested roles. The
+built-in options are:
+
+| ID | Domain |
+|---|---|
+| `software` | Software development |
+| `healthcare` | Healthcare and medicine |
+| `legal` | Legal services |
+| `finance` | Finance and operations |
+| `marketing` | Marketing and growth |
+| `sales` | Sales and customer success |
+| `research` | Research and education |
+| `operations` | Operations and delivery |
+| `custom` | Another domain without assuming an industry |
+
+### 2. Seat role
+
+This is the responsibility of the city lead. It is not the city name or runtime.
+The seat remains chair even when you choose `blank`.
+
+### 3. Repositories and each agent's role
+
+You can read repositories from disk, your GitHub account, or a GitHub
+organisation. Each selected local repo receives:
+
+- a tmux window;
+- a private bus actor;
+- its repository as working directory;
+- an explicit operating role;
+- the skills its runtime can already discover there.
+
+A repo role may come from another domain. For example, a software city can give
+`po` to its product repo, `seo` to a portfolio, and `data-engineer` to a
+pipeline. None becomes chair.
+
+Selecting zero repos is also valid: the city opens with only its seat.
+
+### 4. Goal
+
+A goal may be quantitative or qualitative. It stores:
+
+- title;
+- observed signal;
+- a command that returns the measure, when one exists;
+- person and review frequency for a qualitative judgement;
+- baseline;
+- target;
+- target date.
+
+You may skip it and configure it later with `agents-city seat --goal`.
+
+### 5. Runtime for each window
+
+Pressing Enter keeps Claude everywhere. You may instead choose per window:
+
+- Claude and, optionally, model/effort;
+- Codex;
+- OpenCode;
+- Kimi;
+- an unknown command through the explicit terminal fallback.
+
+Persistent choices live in the owner card. `seat --model` and `--effort` are
+one-launch overrides only.
+
+### What gets created
+
+```text
+~/.agents-city/
+├── .runtime/                  # bus endpoints, queues, and ephemeral state
+├── state/                     # local map state, separated per city
+├── .backups/                  # old migrations
+└── <owner>/
+    ├── .current                # selected city
+    ├── .backups/              # recoverable owner resets
+    └── <city>/
+        ├── city.yml            # id, name, slug, owner, domain
+        ├── roads.json          # allowed roads
+        ├── <owner>.md          # role, repos, roles, goal, runtimes
+        ├── AGENTS.md           # how to read this city
+        ├── domains/            # editable domain knowledge
+        ├── roles/              # editable role knowledge
+        ├── deliberations/      # committee state, events, and acts
+        ├── units.yml           # map districts, when used
+        └── parcels.yml         # map houses/parcels, when used
+```
+
+The `~/.agents-city` root is a container, never a city. `home` is simply the
+first city and is isolated exactly like `product` or `client-a`.
+
+## Working inside tmux
+
+A session is named `<owner>-<city>` and contains:
+
+- a `seat` window, located in the city folder;
+- one window per locally found repo;
+- the configured runtime already started in each window.
+
+Shortcuts are added only to the currently running tmux server:
+
+| Action | Shortcut |
+|---|---|
+| Switch to windows 1–9 | `Alt+1` … `Alt+9` |
+| Previous/next window | `Alt+←` / `Alt+→` |
+| Select with mouse | click the bottom status bar |
+| Scroll | mouse wheel |
+| Detach without closing | `Ctrl-b`, then `d` |
+| Return to the city | `agents-city seat --city <name>` |
+
+If the session already exists, `seat` reattaches to it; it does not create a
+duplicate set of agents. A bell or activity-coloured tab means that window may
+need attention.
+
+Claude starts in a stagger because multiple instances share its OAuth token.
+Codex, OpenCode, and Kimi do not wait for that stagger. Use
+`CITY_SETTLE=0 CITY_STAGGER=0` only when you deliberately want to disable it.
+
+Do not close a city by killing generic processes. Use:
+
+```bash
+agents-city exit <city> --dry-run
+agents-city exit <city>
+```
+
+## Runtimes and transports
+
+All runtimes receive typed envelopes from the same local WebSocket bus, but each
+provider has a native last mile:
+
+| Runtime | Bus task delivery | Visible interface | Requirement |
+|---|---|---|---|
+| Claude | persistent `stream-json` over stdin/stdout | interactive `city>` gateway plus visible Claude transcript | authenticated `claude` CLI; no Team account or admin policy |
+| Codex | `app-server` WebSocket | official TUI connected with `codex --remote` | authenticated `codex` CLI |
+| OpenCode | HTTP/SSE API | interactive gateway `city>` console | configured `opencode` CLI |
+| Kimi | REST + WebSocket | interactive gateway `city>` console | configured `kimi` or `kimi-code` CLI |
+| Unknown CLI | compatibility adapter | its own TUI/command inside tmux | explicit `terminal:<command>` setting |
+
+Claude, Codex, OpenCode, and Kimi tasks are **not** pasted into tmux and do not
+use the clipboard. Terminal fallback exists only for an explicitly selected
+unknown command.
+
+Agents City does **not** use custom Claude Channels in its normal launch path.
+The same official Claude Code process stays open in print/streaming mode and
+receives JSONL turns directly from the city gateway. A personal Pro/Max account
+therefore needs no `sudo`, `managed-settings.json`, Team console, Channel
+allowlist, development bypass, or per-window confirmation. The installed plugin
+still supplies its MCP tools, skills and hooks normally. Custom Channels remain
+an optional upstream preview mechanism, not a prerequisite for Agents City.
+
+Conceptual multi-model card settings:
+
+```yaml
+runs.seat: codex
+runs.api: codex --model gpt-5
+runs.analytics: opencode -m lmstudio/qwen3-coder
+runs.research: kimi
+runs.legacy: terminal:gemini
+model.docs: sonnet
+effort.docs: high
+```
+
+When the seat does not use Claude, it has no `/city:` commands. Everything
+fundamental remains available through `agents-city committee`, `road`, `bus`,
+`skills`, `seat`, `reset`, and `exit`.
+
+## Domains, roles, and knowledge
+
+### Built-in roles by domain
+
+| Domain | Available role IDs |
+|---|---|
+| `software` | `cpto`, `dev`, `data-engineer`, `devops`, `data`, `product-design`, `po`, `llm-engineer`, `ai-manager`, `blank` |
+| `healthcare` | `clinical-director`, `clinician`, `patient-safety`, `clinical-ops`, `health-data`, `health-compliance`, `blank` |
+| `legal` | `managing-partner`, `associate`, `compliance`, `knowledge`, `ops`, `blank` |
+| `finance` | `cfo`, `controller`, `fin-analytics`, `ops`, `compliance`, `blank` |
+| `marketing` | `brand-lead`, `content`, `performance`, `seo`, `lifecycle`, `data`, `product-design`, `blank` |
+| `sales` | `revenue-lead`, `account-executive`, `revops`, `customer-success`, `enablement`, `blank` |
+| `research` | `research-director`, `researcher`, `methods`, `research-ops`, `ethics`, `knowledge`, `blank` |
+| `operations` | `operations-lead`, `program-manager`, `process-owner`, `quality`, `knowledge`, `blank` |
+| `custom` | `city-lead`, `specialist`, `quality`, `knowledge`, `blank` |
+
+`blank` is a complete choice: it creates no role file, applies no hidden
+profile, and infers no responsibility. You may assign it to the seat or any repo
+and change it later.
+
+When you select a domain/role, Agents City copies initial packs into the city:
+
+```text
+domains/<domain>.md
+roles/<role>.md
+```
+
+They are ordinary Markdown. You can edit, remove, or extend their contents. A
+later role change does not overwrite an existing file, so your adaptations are
+preserved. This knowledge is not a skill.
+
+## Complete command reference
+
+### Overview
+
+```text
+agents-city [hall]
+agents-city setup
+agents-city seat
+agents-city cities
+agents-city road
+agents-city bus
+agents-city committee
+agents-city skills
+agents-city city
+agents-city demo
+agents-city report
+agents-city tokens
+agents-city logs
+agents-city benchmark
+agents-city reset
+agents-city exit
+agents-city test
+```
+
+Global commands:
+
+```bash
+agents-city --help
+agents-city --version
+```
+
+### `agents-city` and `agents-city hall`
+
+Open the local Hall for the selected city.
+
+```bash
+agents-city
+agents-city hall
+agents-city hall --city product
+agents-city hall --no-browser
+```
+
+| Option | Effect |
+|---|---|
+| `--city NAME|ID|PATH` | select a known city and open it |
+| `--no-browser` | do not open the browser; print the local URL and temporary token |
+
+The server binds only to `127.0.0.1` and requires a per-run token for writes.
+The **City live** column on the right connects as a spectator to the same local
+WebSocket bus used by the agents. It shows ordinary visible user/agent messages,
+runtime failures, and the complete moderated committee flow as a conversation:
+one avatar per repo, the seat marked as chair, and one visible turn per revealed
+position or granted reply. Routine commands and lifecycle noise are collapsed
+behind **show work**; the selected conversation opens by default.
+
+The Hall opens directly on **The map**. The city owns the whole centre canvas: state, controls, and
+history stay in the side rails, never above or below the map. Every semantic
+turn arriving on that same WebSocket also creates a short game-style bubble,
+anchored to the speaking character and prefixed with its recipient (`Para
+seat:`, `Para committee:`, and so on). The bubble is only the transient summary;
+the complete message and evidence remain in **City live**. No decorative
+dialogue, commands, private reasoning, or raw envelopes become speech.
+
+Codex uses completed visible app-server items; Claude uses its documented prompt
+and stop hooks. Provider reasoning items, chain-of-thought, credentials, and raw
+transport frames are neither shown nor written to the activity log. The
+spectator token rotates with the hub, accepts only an origin on this computer,
+and is read-only: the browser cannot direct the committee. `Ctrl-c` stops the
+Hall.
+
+### `agents-city setup`
+
+Creates or selects a city and opens the Hall; `--tui` hands the flow to `seat`.
+
+```bash
+agents-city setup
+agents-city setup --city product
+agents-city setup --city product --tui
+agents-city setup --out /path/to/a/city
+agents-city setup --demo
+agents-city setup --no-browser
+```
+
+| Option | Effect |
+|---|---|
+| `--city NAME` | create the managed city when missing or select the existing one |
+| `--out PATH` | register/import an explicit compatible folder; advanced use |
+| `--demo` | open the complete guided Aurora Games demo |
+| `--tui` | use terminal onboarding and open the session |
+| `--no-browser` | keep the Hall in the terminal and print its URL |
+
+### `agents-city seat`
+
+Configures requested settings, ensures tmux/plugin, and opens or resumes the city
+session.
+
+```bash
+agents-city seat
+agents-city seat --city product
+agents-city seat --repos
+agents-city seat --agent-roles
+agents-city seat --goal
+agents-city seat --engines
+agents-city seat --domain marketing
+agents-city seat --domain marketing --role brand-lead
+agents-city seat --role blank
+agents-city seat --only api,web
+agents-city seat --model sonnet --effort high
+agents-city seat --seat-yolo on
+agents-city seat --no-yolo --no-sync
+```
+
+| Option | Persistence and effect |
+|---|---|
+| `--city NAME|PATH` | select this city and open its session |
+| `--repos` | choose repos again, then their agent roles; persists |
+| `--agent-roles`, `--agents` | choose only each repo role again; persists |
+| `--goal` | redefine the goal; persists |
+| `--engines` | choose runtime/model per window; persists |
+| `--domain DOMAIN` | change domain; persists and asks for a compatible role unless `--role` is supplied |
+| `--role ROLE` | change the seat role without a picker; persists |
+| `--only a,b` | open only those repos this run; does not alter the card |
+| `--model ALIAS` | model override for all windows in this launch |
+| `--effort LEVEL` | `low`, `medium`, `high`, `xhigh`, or `max` override for this launch |
+| `--seat-yolo on\|off` | whether the chair itself runs without permission prompts; persists per city (`city.yml seat_yolo`, also question six of the wizard). Locally the seat is the owner's own hands; repo windows keep their own yolo/cage story either way |
+| `--no-yolo` | disable auto-approval for this launch — seat included, whatever `seat_yolo` says |
+| `--no-sync` | skip initial `git fetch/pull` in repos for this launch |
+
+`seat` accepts a positional user for compatibility, but only when it matches the
+resolved local owner. Use `--city` for another city belonging to the same user.
+
+### `agents-city cities`
+
+Manages the local catalogue. Creating or selecting does not start tmux.
+
+```bash
+agents-city cities list
+agents-city cities current
+agents-city cities create product
+agents-city cities use product
+agents-city cities use /path/to/city
+```
+
+| Subcommand | Output/effect |
+|---|---|
+| `list` | known cities; `*` marks the selected one |
+| `current` | absolute path of the selected city |
+| `create NAME` | create `~/.agents-city/<owner>/<slug>/` and select it |
+| `use NAME|PATH` | select an existing city without starting it |
+
+### `agents-city road`
+
+Opens and closes the allowlist of connections between seats.
+
+```bash
+agents-city road list product
+agents-city road connect product client-a
+agents-city road invite product
+agents-city road invite product > product.invitation.json
+agents-city road connect product research.invitation.json
+agents-city road disconnect product client-a
+agents-city road disconnect product <remote-city-id>
+```
+
+| Subcommand | Effect |
+|---|---|
+| `list CITY` | show destination, address, and local/remote status |
+| `connect A B` | when B is local, write both ends symmetrically |
+| `connect A invitation.json` | add only the local end of a remote road |
+| `invite CITY` | print public JSON without a token |
+| `disconnect A B|ID` | remove both local ends or the specified remote ID |
+
+A city cannot connect to itself. Each machine must independently accept the
+other remote invitation.
+
+### `agents-city bus`
+
+Operates messages between seats over declared roads.
+
+```bash
+agents-city bus roster
+agents-city bus inbox
+agents-city bus send alice/research "Please confirm the event X contract"
+agents-city bus send '*' "Notice for every connected city"
+```
+
+| Subcommand | Effect |
+|---|---|
+| `roster` | return roads and known online presence |
+| `inbox` | return and consume pending inbox; append-only history remains |
+| `send owner/city TEXT` | send to one allowed destination |
+| `send '*' TEXT` | send to all roads; requires at least one |
+
+Only `seat` may run these operations. A repo actor is rejected by the ACL even
+if it knows the destination address.
+
+### `agents-city committee`
+
+Manages structured deliberations inside one city. Every command accepts fields
+as flags or as a JSON object through `--input`.
+
+```bash
+agents-city committee list
+agents-city committee history
+agents-city committee show <deliberation-id>
+agents-city committee status <deliberation-id>   # alias for show
+agents-city committee schema open
+agents-city committee open --input proposal.json
+agents-city committee open --input - < proposal.json
+```
+
+| Subcommand | Allowed actor | Purpose |
+|---|---|---|
+| `list` | any involved city actor | open deliberations visible to that actor |
+| `history` | seat | finished decision history and contributor counts |
+| `show ID`, `status ID` | involved city actor | complete actor-visible state and events |
+| `schema VERB` | anyone | JSON contract for a mutation verb |
+| `open` | seat | state question, outcome, members, and boundaries |
+| `respond` | invited member | record one independent initial position |
+| `synthesize` | seat | publish agreements, conflicts, and unknowns |
+| `floor-request` | member | request a turn for evidence, contradiction, risk, or dependency |
+| `floor-grant` | seat | grant one floor request |
+| `floor-deny` | seat | deny one floor request with a reason |
+| `reply` | member holding the floor | submit one bounded, evidence-based reply |
+| `decide` | seat | record outcome, owners, verifier, and reopen conditions |
+| `verify` | assigned verifier | return `pass` or `fail` with checks |
+| `replan` | seat | reopen a failed verification with a new plan |
+| `close` | seat | close an already verified outcome |
+| `cancel` | seat | cancel a deliberation with a reason |
+
+Opening with flags:
+
+```bash
+agents-city committee open \
+  --question "Should we ship today?" \
+  --outcome-wanted "A reversible decision with an owner and verification" \
+  --context "The release candidate passed the local suite" \
+  --constraint "Do not lose data" \
+  --constraint "Rollback within ten minutes" \
+  --done "The decision names an executor and verifier" \
+  --authority execute \
+  --member api \
+  --member web \
+  --member qa \
+  --max-rebuttals 1
+```
+
+The result prints a `deliberationId`. Keep it for later transitions. JSON is
+usually clearer for large payloads:
+
+```json
+{
+  "question": "Should we ship today?",
+  "desiredOutcome": "A reversible decision with an owner and verification",
+  "context": "The release candidate passed the local suite",
+  "constraints": ["Do not lose data", "Rollback within ten minutes"],
+  "definitionOfDone": ["Executor and verifier assigned"],
+  "authority": "execute",
+  "participants": ["api", "web", "qa"],
+  "maxRebuttals": 1
+}
+```
+
+```bash
+agents-city committee open --input proposal.json
+```
+
+`--input -` reads stdin. When JSON and flags are mixed, explicit flags override
+the equivalent field. Repeatable flags are `--member`, `--constraint`, `--done`,
+`--evidence`, `--risk`, `--unknown`, `--agreement`, `--conflict`, `--check`,
+`--residual-risk`, `--selected-evidence`, `--rejected-option`, `--dissent`,
+`--reopen-if`, `--learning`, and `--followup`.
+
+Always inspect the exact contract shipped by the installed version:
+
+```bash
+agents-city committee schema respond
+agents-city committee schema decide
+agents-city committee schema verify
+```
+
+Member commands (`respond`, `floor-request`, and `reply`) are normally executed
+by the authenticated repo agent after receiving an envelope. Running one from
+the seat correctly fails its ACL: accepting an actor name as text would not let
+the bus pretend to be that actor.
+
+### `agents-city skills`
+
+Lists skills already present in a city's repos. This is read-only: it does not
+install, copy, enable, or remove anything.
+
+```bash
+agents-city skills
+agents-city skills product
+```
+
+Recognised layouts per repository:
+
+```text
+SKILL.md
+.claude/skills/*/SKILL.md
+.codex/skills/*/SKILL.md
+.agents/skills/*/SKILL.md
+skills/*/SKILL.md
+```
+
+Actual invocation depends on the runtime. Agents City advertises the capability
+for the member and lets the provider enforce its own discovery and use rules.
+
+### `agents-city city`
+
+Opens one city's local map without starting an agent session.
+
+```bash
+agents-city city
+agents-city city ~/.agents-city/alice/product
+```
+
+It uses port `8787` or the next free port, binds to loopback, and opens the
+browser. `Ctrl-c` stops the server. `units.yml`, `parcels.yml`, the card, and bus
+state feed the visualisation.
+
+The map is live, not a postcard. Three layers stage what is happening right
+now, all derived from data the product already emits: presence (a mid-turn
+house glows and breathes, a stopped one cools), the town hall (committee
+sessions play on stage — sealed positions fly in face down, the floor is a
+raised hand, verification stamps the door, closing files the act — with the
+camera flying to the session and members walking over), and one gate per road,
+which letters to other cities leave through. Agents get deterministic
+identicon faces, `knowledge`/`coordinator` parcels wear a different building
+family than `code`, the town hall and the gates are clickable, `P` (or the ⛶
+control) toggles fullscreen, and the Hall's live rail is resizable by dragging
+its edge. The full contract is in
+[docs/map-live-layers.md](docs/map-live-layers.md).
+
+### `agents-city demo`
+
+Opens one fictional, disposable demo city in the complete Hall. The map owns
+the centre; the right rail plays a guided deliberation and those same turns
+appear as `Para …:` speech bubbles over their agents. There is one demo per
+domain — real chaos told in plain words, not programmer phrases:
+
+```bash
+agents-city demo                     # software · Aurora Games — the night the saves vanished
+agents-city demo --domain medicina   # Clínica Alba — the morning the appointments doubled
+agents-city demo --domain legal      # Costa & Ley — the deadline at nine tomorrow
+agents-city demo --no-browser
+```
+
+It starts no models and needs no Claude, Codex, OpenCode or Kimi account. The
+stories are declared presentation content, but their engineering is not an
+animation: all 22 events cross the real authenticated WebSocket, committee
+state machine, durable ledger and spectator feed. Every story walks the WHOLE
+machine, including the part demos usually hide: three isolated positions, two
+chair-granted floor requests, a decision, a verification that FAILS, a replan,
+and only then a verified close. The clinic and the firm are agents-first
+cities — knowledge and coordinator agents, no repositories — so they also
+exercise the roster and the map's building families.
+
+The Hall's live rail shows a framed **guided committee** control for demo
+cities only: `⟳ replay` plays the domain's story again, and `⏸ pause` /
+`▶ resume` stop and continue the storyteller process itself (`SIGSTOP`, a real
+pause). `/api/demo` refuses any city that is not a packaged demo: a real
+city's committee is real, and a replay there would publish fiction onto a real
+bus.
+
+The demo copies its city and runtime into a temporary directory. `Ctrl-c` stops
+its Hall, map and hub and removes that copy; it never selects, rewrites or starts
+your cities. If another map already owns `8787`, the demo uses another port and
+the Hall checks city identity instead of accidentally framing the wrong map.
+
+### `agents-city report`
+
+Computes growth that can be represented on the map and optionally pushes it to
+the configured city service.
+
+```bash
+agents-city report
+agents-city report --data ~/.agents-city/alice/product
+agents-city report --url https://city.example.com --token "$CITY_TOKEN"
+agents-city report --push --quiet
+```
+
+| Option | Effect |
+|---|---|
+| `--data PATH` | use another city data folder |
+| `--url URL` | override the service URL |
+| `--token TOKEN` | override the authentication token |
+| `--push` | send the report; without it, only compute/show |
+| `--quiet` | reduce human-readable output |
+
+### `agents-city tokens`
+
+Aggregates local Claude transcript usage and can send totals only. It does not
+send prompts, responses, or file paths.
+
+```bash
+agents-city tokens
+agents-city tokens --days 7
+agents-city tokens --all
+agents-city tokens --push --quiet
+agents-city tokens --url https://city.example.com --token "$CITY_TOKEN"
+```
+
+| Option | Effect |
+|---|---|
+| `--days N` | time window; default 30 days |
+| `--all` | re-read transcripts inside `--days`, ignoring the incremental cache |
+| `--url URL` | override the service URL |
+| `--token TOKEN` | override the authentication token |
+| `--push` | send aggregates; without it, only show them |
+| `--quiet` | reduce human-readable output |
+
+`tokens` does not automatically estimate Codex, OpenCode, or Kimi usage.
+
+### `agents-city logs`
+
+Reads the selected city's two durable local streams: visible semantic activity
+and secret-scrubbed operational diagnostics. It does not read provider reasoning.
+
+```bash
+agents-city logs
+agents-city logs --activity --lines 50
+agents-city logs --diagnostics --lines 200
+agents-city logs --follow
+agents-city logs --json --follow
+```
+
+| Option | Effect |
+|---|---|
+| `--activity` | only visible prompts, answers, work, and committee events |
+| `--diagnostics` | only hub, socket, gateway, hook, and launcher diagnostics |
+| `-n, --lines N` | initial number of records; default 100 |
+| `-f, --follow` | continue streaming appended records until `Ctrl-c` |
+| `--json` | emit the stored JSONL records unchanged |
+
+The files live under the selected city's private runtime directory as
+`activity.jsonl` and `diagnostics.jsonl`. They survive a Hall reload and a bus
+restart, are mode `0600`, and can be inspected directly. Activity source IDs
+make repeated provider notifications and hooks idempotent.
+
+### `agents-city benchmark`
+
+Measures transport, real runtimes, or the structure of the committee protocol.
+
+#### Local stress without model quota
+
+```bash
+agents-city benchmark stress
+agents-city benchmark stress --agents 40 --rounds 2 --timeout 20
+agents-city benchmark stress --agents 80 --rounds 5 --json
+agents-city benchmark stress --keep
+```
+
+| Option | Effect |
+|---|---|
+| `--agents N` | simulated actors; must be even, default 40 |
+| `--rounds N` | rounds per actor; default 2 |
+| `--timeout SEC` | benchmark limit; default 20 |
+| `--json` | machine-readable output |
+| `--keep` | preserve the temporary workspace for inspection |
+
+#### Real runtimes, consuming quota
+
+```bash
+agents-city benchmark live --runtime claude --runtime codex
+agents-city benchmark live \
+  --runtime codex \
+  --runtime kimi \
+  --timeout 180 \
+  --json
+agents-city benchmark live \
+  --command codex="codex --model gpt-5" \
+  --command opencode="opencode -m lmstudio/qwen3-coder" \
+  --keep
+```
+
+| Option | Effect |
+|---|---|
+| `--runtime RUNTIME` | runtime to measure; repeatable: `claude`, `codex`, `kimi`, `opencode` |
+| `--command RUNTIME=COMMAND` | concrete command for that runtime; repeatable |
+| `--timeout SEC` | limit per case; default 180 |
+| `--json` | machine-readable output |
+| `--no-save` | do not append the result to local history |
+| `--keep` | preserve temporary workspaces |
+
+`live` makes real calls to installed providers and can consume quota or money.
+Check authentication and limits before running it.
+
+#### Committee protocol
+
+```bash
+agents-city benchmark committee
+agents-city benchmark committee --json
+```
+
+This compares the structured flow with an unbounded chat: response barrier,
+floor control, decision, and verification. It is a deterministic structural
+benchmark; by itself it proves neither higher answer quality nor a SOTA claim.
+
+### `agents-city reset`
+
+Resets **one** managed city while preserving its stable identity and repositories.
+
+```bash
+agents-city reset product --dry-run
+agents-city reset product
+```
+
+The reset plan:
+
+1. verifies the target is a managed city, not an arbitrary path;
+2. shows and stops only that city's session/runtime;
+3. creates a recoverable backup under its owner;
+4. preserves `id`, owner, name, and slug;
+5. removes that city's configuration, deliberations, and generated state;
+6. does not touch source repositories;
+7. symmetrically removes incident local roads;
+8. leaves the city ready for onboarding again.
+
+There is no automatic `restore` command yet. The exact backup path is printed
+for manual recovery. Always run `--dry-run` first.
+
+### `agents-city exit`
+
+Stops Agents City sessions and processes without deleting configuration.
+
+```bash
+agents-city exit product --dry-run
+agents-city exit product
+agents-city exit --dry-run
+agents-city exit
+```
+
+With a city, it closes only that city's tmux, gateway, and helper processes; the
+Hall may stay up. Without a city, it shows or closes everything managed by
+Agents City. A tmux session may contain unsaved work, so dry-run is the safe way
+to confirm scope.
+
+### `agents-city test`
+
+Runs checkout tests. With no arguments it runs every suite; with names it runs
+only those suites.
+
+```bash
+agents-city test
+agents-city test seat runtime-ui
+agents-city test committee stress benchmark
+```
+
+Available suites:
+
+```text
+widgets card parcels domains serve seat cities channel committee live-feed
+runtime runtime-ui runtime-failures stress adapter benchmark contracts exit
+cage broker launch
+```
+
+This command is intended for contributors or local tarball validation. Normal
+use does not require running tests at every startup.
+
+## Committee: complete workflow
+
+The committee behaves like a management committee: the seat frames and chairs
+the decision; specialists contribute evidence from their repos; nobody opens a
+lateral conversation; the seat integrates and another identity verifies.
+
+```text
+open
+  └─ collecting: independent, hidden positions
+       ├─ missing replies + proceedWithout ─┐
+       └─ all reply -> review              │
+                                             v
+                                       synthesize
+                                             │
+                                             v
+                                       deliberating
+                                ┌─ bounded floor ─┐
+                                └─ request/reply ┘
+                                             │
+                                           decide
+                                             │
+                                             v
+                                         verifying
+                                   ┌─ fail ─└─ pass
+                                   v               v
+                          verification_failed   verified
+                                   │               │
+                                 replan           close
+                                   │               │
+                                   └─> review        closed
+```
+
+### 1. Prepare the brief
+
+A good question names a decision, not merely a topic. The desired outcome says
+what the committee must produce; `definitionOfDone` lists observable conditions.
+Select only repositories capable of producing relevant evidence.
+
+| `open` field | Required | Values/meaning |
+|---|---|---|
+| `question` | yes | exact decision |
+| `desiredOutcome` | yes | concrete expected result |
+| `context` | no | minimum necessary facts |
+| `constraints` | no | time, cost, security, or policy boundaries |
+| `definitionOfDone` | yes, list | observable acceptance conditions |
+| `authority` | no | `recommend`, `decide`, or `execute`; default `recommend` |
+| `participants` | yes, list | repo actor names in this city |
+| `maxRebuttals` | no | integer 0–5; default 2 per member |
+
+`authority` records the mandate; it does not change technical ACLs.
+
+### 2. Collect isolated positions
+
+Each participant receives the same brief and responds once:
+
+```bash
+agents-city committee respond "$DELIBERATION_ID" \
+  --stance conditional \
+  --recommendation "Ship to 10% first" \
+  --evidence "npm test: 844 checks passed" \
+  --expected-impact "Detect regressions before full rollout" \
+  --visible-when "After 30 minutes of telemetry" \
+  --withdraw-if "The migration is not reversible" \
+  --risk "Insufficient canary capacity" \
+  --unknown "First-hour production load"
+```
+
+`stance` is `support`, `oppose`, `conditional`, or `abstain`. `evidence` is
+required and repeatable. The runtime executes the response inside the repo
+window under its real identity. Until the barrier opens, the seat sees progress,
+not initial-position contents; this reduces anchoring.
+
+### 3. Synthesise without voting
+
+Once all positions are ready, the seat integrates evidence:
+
+```bash
+agents-city committee synthesize "$DELIBERATION_ID" \
+  --summary "There is agreement on a reversible canary" \
+  --agreement "The migration needs a tested rollback" \
+  --conflict "10% versus 25% initial traffic" \
+  --unknown "Capacity at expected peak"
+```
+
+If a member is missing, it cannot simply be ignored:
+
+```bash
+agents-city committee synthesize "$DELIBERATION_ID" \
+  --summary "Provisional synthesis" \
+  --proceed-without "QA is offline; the deadline is today and rollback remains available"
+```
+
+The decision integrates evidence, impact, and withdrawal conditions rather than
+counting votes.
+
+### 4. Request and grant the floor
+
+After synthesis, a member may reply only with an admitted basis:
+
+```bash
+agents-city committee floor-request "$DELIBERATION_ID" \
+  --basis new_evidence \
+  --reason "The canary failed its rollback test" \
+  --evidence "artifacts/rollback.log: exit 1"
+```
+
+`basis` accepts `new_evidence`, `contradiction`, `risk`, or `dependency`. The
+seat resolves the returned `requestId`:
+
+```bash
+agents-city committee floor-grant "$DELIBERATION_ID" --request-id "$REQUEST_ID"
+# or:
+agents-city committee floor-deny "$DELIBERATION_ID" \
+  --request-id "$REQUEST_ID" \
+  --reason "The evidence is already in the synthesis"
+```
+
+Once granted, that member has exactly one reply and releases the floor by using it:
+
+```bash
+agents-city committee reply "$DELIBERATION_ID" \
+  --claim "Shipping with the current script is unsafe" \
+  --evidence "artifacts/rollback.log: exit 1" \
+  --consequence "Block until fixed and rerun rollback"
+```
+
+The reply reaches the seat **and is heard by the whole committee**. Other members
+do not answer the speaker directly: if one finds new evidence, a contradiction,
+a material risk, or a dependency, it asks the seat for another turn. The seat
+grants or denies it, and only then may that agent speak. This is real specialist
+conversation mediated like an executive committee, not an all-to-all chat. Two
+active turns cannot coexist; every grant permits one intervention;
+`maxRebuttals` bounds the cascade per member; and the seat must resolve every
+pending request before deciding.
+
+### 5. Decide and attribute
+
+```bash
+agents-city committee decide "$DELIBERATION_ID" \
+  --outcome "Fix rollback and ship a 10% canary" \
+  --rationale "This limits impact and satisfies reversibility" \
+  --owner "Release owner" \
+  --executor api \
+  --verifier qa \
+  --verification-question "Do rollback and canary pass end to end?" \
+  --selected-evidence "full suite is green" \
+  --selected-evidence "reproducible rollback failure" \
+  --decisive-contributors qa \
+  --rejected-option "Immediate full rollout" \
+  --dissent "web prefers a 25% canary" \
+  --reopen-if "5xx errors exceed 1% for five minutes"
+```
+
+`selectedEvidence`, `decisiveContributors`, and `reopenIf` are required. If
+another identity is available, `verifier` cannot equal `executor`. Dissent stays
+in the act even when it does not change the decision. Use JSON input when more
+than one decisive contributor must be recorded.
+
+### 6. Verify, replan, or close
+
+Only the assigned verifier may run:
+
+```bash
+agents-city committee verify "$DELIBERATION_ID" \
+  --result pass \
+  --evidence "artifacts/e2e-rollback.txt" \
+  --check "canary returns 200" \
+  --check "rollback restores the previous version" \
+  --residual-risk "The first hour at full load remains unobserved"
+```
+
+After `fail`, the seat must replan and then synthesise/decide again:
+
+```bash
+agents-city committee replan "$DELIBERATION_ID" \
+  --reason "Rollback still leaves an incompatible schema"
+```
+
+After `pass`, the seat may close:
+
+```bash
+agents-city committee close "$DELIBERATION_ID" \
+  --summary "Canary verified; rollout authorised" \
+  --learning "Test rollback before fixing a release window" \
+  --followup "Watch 5xx rates for the first hour"
+```
+
+A deliberation cannot close without reproducible passing verification. If it is
+no longer relevant, the seat may cancel it:
+
+```bash
+agents-city committee cancel "$DELIBERATION_ID" \
+  --reason "The release was replaced by another candidate"
+```
+
+State, events, and a readable act remain in `deliberations/`. `history`
+summarises recent decisions and decisive contributions to expose repeated
+influence. That count is a review signal, not automatic evidence of capture.
+
+## Claude `/city:` commands
+
+These commands come from the Claude plugin. They do not exist inside Codex,
+OpenCode, or Kimi TUIs; use the equivalent `agents-city` terminal commands there.
+
+| Command | Use case |
+|---|---|
+| `/city:setup [--city N] [--tui] [--demo]` | create/open a city through the shared flow |
+| `/city:join [--domain D\|--role R\|--repos\|--agent-roles\|--goal\|--engines]` | compatibility name for configuring the seat; does not add a person |
+| `/city:session [--no-yolo] [--only a,b]` | open or resume this city's tmux |
+| `/city:settings [domain\|role\|repos\|agent-roles\|goal\|engines\|roads\|skills]` | read or change one configuration area |
+| `/city:goals` | show or edit the current goal |
+| `/city:committee QUESTION` | prepare and open a chaired deliberation |
+| `/city:committee status ID` | inspect the next legal transition |
+| `/city:round [--to owner/city] [--since DATE]` | compare goal and local evidence; consult relevant roads |
+| `/city:notice [--pr N\|--since REF] [--dry]` | notify only affected cities about a verified change |
+| `/city:propose owner/city [SUBJECT]` | send an evidence-backed proposal |
+| `/city:team` | historical alias: list cities, active city, repos, and roads; not people |
+| `/city:exit [CITY] [--dry-run]` | show or close managed processes |
+
+`/city:notice --dry` sends nothing. `/city:round` and `/city:propose` can only use
+destinations present in `road list`. A reply from another city informs the seat;
+it never gains authority to command a local repo directly.
+
+## Use-case cookbook
+
+### Case 1: start from zero with one city and Claude
+
+```bash
+cd /path/to/agents-city
+npm pack
+npm install -g ./agents-city-0.3.0-beta.21.tgz
+agents-city seat
+```
+
+1. Choose the domain.
+2. Choose the seat role.
+3. Select repositories or continue with none.
+4. Define or skip the goal.
+5. Press Enter in the runtime picker to keep Claude.
+
+Result: a `home` city, an `<owner>-home` session, one `seat` window, and one
+window per selected local repo. Agents City keeps one official Claude Code
+process per window and feeds it through persistent `stream-json`; it does not
+request a custom Channel or require admin/per-window approval.
+
+### Case 2: use Codex as the main seat
+
+```bash
+agents-city seat --engines
+```
+
+Choose Codex on the `seat` row, confirm the other rows, and open the city. Agents
+City starts `codex app-server` on loopback and opens the official TUI with
+`codex --remote`. The TUI creates its persisted thread; the gateway detects only
+the new thread for that working directory and joins it through `thread/resume`.
+You should type directly into Codex. A `city>` prompt in a Codex seat means an
+old version or a failed launch; it is not the intended Codex interface.
+
+### Case 3: mix runtimes per repository
+
+```bash
+agents-city seat --city product --engines
+```
+
+Example selection:
+
+```text
+seat       Codex
+api        Claude / opus model / high effort
+web        Codex
+analytics  OpenCode
+research   Kimi
+```
+
+Each choice persists in the card. The next `seat` run reuses it. To test another
+combination without retaining old processes:
+
+```bash
+agents-city exit product --dry-run
+agents-city exit product
+agents-city seat --city product --engines
+```
+
+### Case 4: use a local model through OpenCode
+
+Agents City does not choose OpenCode's provider. In the runtime picker, select
+OpenCode and enter the command/model accepted by your installation, for example:
+
+```text
+opencode -m lmstudio/qwen3-coder
+```
+
+Validate it independently first:
+
+```bash
+opencode -m lmstudio/qwen3-coder
+```
+
+Then run `agents-city seat --engines`. Bus delivery reaches OpenCode through
+HTTP/SSE; the model may be local while Agents City keeps the same typed envelope.
+
+### Case 5: use a CLI that is not natively integrated yet
+
+Choose “another command (terminal fallback)” under `--engines` and enter, for
+example, `gemini`. Agents City stores:
+
+```yaml
+runs.api: terminal:gemini
+```
+
+The prefix makes it explicit that this window may need visible tmux injection.
+An unknown command without `terminal:` is rejected when reading a hand-edited
+card; known runtimes never silently degrade to terminal transport.
+
+### Case 6: create several cities for the same user
+
+```bash
+agents-city cities create product
+agents-city seat --city product
+
+agents-city cities create client-a
+agents-city seat --city client-a
+
+agents-city cities list
+```
+
+Expected result:
+
+```text
+~/.agents-city/<owner>/product/
+~/.agents-city/<owner>/client-a/
+```
+
+Each city has its own domain, role, goal, repos, recognised skills,
+deliberations, roads, runtime, and tmux session. `home` has no special privilege.
+
+### Case 7: give every repo a different speciality
+
+```bash
+agents-city seat --city product --agent-roles
+```
+
+Assign `po` to the main repo, `seo` to the portfolio, and `data-engineer` to the
+pipeline even when the seat domain is `software`. The picker can search roles
+from other domains. Speciality changes perspective and editable context; all
+repo agents retain technical `member` authority.
+
+### Case 8: work without a preloaded profile
+
+```bash
+agents-city seat --city lab --role blank
+agents-city seat --city lab --agent-roles
+```
+
+Also select `blank` for repos that should receive no profile. No role knowledge
+file is created and no hidden role is inferred. Repo instructions and skills
+continue to work normally.
+
+### Case 9: connect two local cities
+
+```bash
+agents-city road connect product client-a
+agents-city road list product
+agents-city road list client-a
+```
+
+The road is written at both ends. Start both cities and, from one seat:
+
+```bash
+CITY_OWNER=alice
+AGENTS_CITY_DATA="$HOME/.agents-city/$CITY_OWNER/product" \
+  agents-city bus send "$CITY_OWNER/client-a" "Does this change affect your contract?"
+```
+
+Inside a normal session, you do not need to set `AGENTS_CITY_DATA`; it is already
+injected into each window. The example makes it explicit for an outside terminal.
+
+### Case 10: connect cities belonging to different machines or people
+
+On machine A:
+
+```bash
+agents-city road invite product > product.invitation.json
+```
+
+Transfer that JSON over an appropriate channel. It contains no bus token. On
+machine B:
+
+```bash
+agents-city road connect research product.invitation.json
+agents-city road invite research > research.invitation.json
+```
+
+Return B's invitation and accept it on A:
+
+```bash
+agents-city road connect product research.invitation.json
+```
+
+Both machines need the same compatible remote transport and valid credentials
+through `CITY_BUS_URL`/`CITY_BUS_TOKEN`. Invitations only declare the allowlist;
+they neither deploy infrastructure nor share secrets. See
+[docs/self-host.md](docs/self-host.md) for the remote Worker.
+
+### Case 11: ask several repos for a decision without a group chat
+
+From a Claude seat:
+
+```text
+/city:committee Can we enable the new migration in production?
+```
+
+From any other runtime, prepare the brief and use:
+
+```bash
+agents-city committee open --input migration-decision.json
+```
+
+The seat selects only relevant repos. Initial responses stay isolated; synthesis,
+floor requests, an attributed decision, and verification follow. Use
+`agents-city committee show ID` to inspect state, not to bypass the next legal
+actor.
+
+### Case 12: start only one or two repos in a large city
+
+```bash
+agents-city seat --city product --only api
+agents-city seat --city product --only api,web
+```
+
+`--only` filters windows for this launch; it does not remove repos or roles from
+the card. If a session with another composition already exists, close it first:
+
+```bash
+agents-city exit product --dry-run
+agents-city exit product
+```
+
+### Case 13: select or clone private GitHub repositories
+
+```bash
+agents-city seat --repos
+```
+
+Choose “my GitHub account” or “GitHub organisation”. If `gh` is missing, the
+wizard tries to install it; if unauthenticated, it opens `gh auth login --web`.
+The browser or device code authenticates `gh`, not Agents City. Private repos
+appear only when the token has access. A selected but uncloned repo may stay on
+the card without a window or be cloned, after confirmation, beneath
+`CITY_CODE_DIR` (default `~/codigo`).
+
+### Case 14: inspect skills without installing them
+
+```bash
+agents-city skills product
+```
+
+If `api/.codex/skills/migrations/SKILL.md` appears, the repo already owns that
+skill. Agents City neither copies it into the city nor forces the agent to use
+it. Adding, editing, or removing the `SKILL.md` changes the next read without
+reinstalling Agents City.
+
+### Case 15: open only the map or the guided demo
+
+```bash
+agents-city city ~/.agents-city/<owner>/product
+agents-city demo
+```
+
+`city` renders that city's real data. `demo` opens the complete Aurora Games Hall
+and plays presentation agents over the real infrastructure without invoking
+models. Use `agents-city hall` to manage your own cities.
+
+### Case 16: measure performance and detect regressions
+
+Measure the deterministic bus first, without models:
+
+```bash
+agents-city benchmark stress --agents 40 --rounds 2 --json
+```
+
+Keep the JSON as a baseline. Then, if model quota use is acceptable, measure the
+real path:
+
+```bash
+agents-city benchmark live \
+  --runtime claude \
+  --runtime codex \
+  --runtime kimi \
+  --timeout 180 \
+  --json
+```
+
+Compare bus-to-runtime acceptance separately from end-to-end completion. A
+correct model response does not turn transport delay into “reasoning time”; an
+authentication failure is not counted as a fast sample.
+
+### Case 17: update without leaving old code in live sessions
+
+```bash
+cd /path/to/agents-city
+npm pack
+npm install -g ./agents-city-0.3.0-beta.21.tgz
+agents-city --version
+agents-city exit product --dry-run
+agents-city exit product
+agents-city seat --city product
+```
+
+Installing a tarball does not rewrite live processes. Restarting only the target
+city avoids closing another city or an unrelated tmux session.
+
+### Case 18: clear one city's configuration and repeat onboarding
+
+```bash
+agents-city reset lab --dry-run
+agents-city reset lab
+agents-city seat --city lab
+```
+
+Reset preserves the `lab` identity, creates a backup, and does not touch repos.
+Use `exit`, not `reset`, when you only need to stop processes.
+
+## Files and environment variables
+
+### `city.yml`: city identity
+
+This file stores stable identity and domain. Do not create a city by copying a
+folder and reusing its `id`; use `cities create`.
+
+```yaml
+id: city_a1b2c3d4
+name: product
+slug: product
+owner: alice
+domain: software
+seat_yolo: 1
+```
+
+The public address is derived as `owner/slug`. It is not stored as one global
+plugin identity because several cities may run concurrently. `seat_yolo: 1`
+launches the chair itself without permission prompts — set at the wizard's
+sixth question or with `agents-city seat --seat-yolo on|off`; `--no-yolo`
+still brakes the whole session.
+
+### `<owner>.md`: seat card
+
+This is Markdown with frontmatter. It stores the seat role, repos, each repo
+role, goal, and runtime per window. Reduced example:
+
+```yaml
+---
+user: alice
+name: alice
+role: cpto
+agent: alice-product-cpto
+repos: [api, web, portfolio]
+role.api: data-engineer
+role.web: dev
+role.portfolio: seo
+goals_defined: true
+runs.seat: codex
+runs.api: claude
+model.api: opus
+effort.api: high
+runs.web: codex
+runs.portfolio: terminal:gemini
+---
+```
+
+Names after the dot use the normalised window actor: lowercase letters, digits,
+and hyphens. Prefer `seat --repos`, `--agent-roles`, `--goal`, and `--engines` to
+maintain the card safely. A malformed manual edit degrades the operating role to
+`blank` or may block startup; it is not evaluated as code.
+
+The card body retains the goal and round history. Changing the goal rewrites
+only that section, not the history.
+
+### Editable knowledge
+
+```text
+domains/<domain>.md
+roles/<role>.md
+AGENTS.md
+```
+
+The first two begin as initial profiles and then belong to the city. You may
+edit, replace, or remove them. Agents City does not overwrite an existing file
+during a later configuration change. `AGENTS.md` tells runtimes how to interpret
+the city; review it after deep customisation.
+
+Skills remain inside repositories and are independent of these files.
+
+### Runtime state
+
+The local hub keeps ephemeral state separate from readable configuration:
+
+```text
+~/.agents-city/.runtime/bus/<city-id>/
+├── endpoint.json
+├── hub.lock
+├── road-token
+├── actors/*.json
+├── outbox/<actor>/*.json
+├── road-queue/*.json
+├── road-inbox/*.json
+└── road-history.jsonl
+```
+
+Credentials and runtime files are created with private permissions. Outboxes let
+an actor reconnect without losing an already accepted task; its ACK removes the
+pending item. Current limits are 200 pending items per queue and a 72-hour
+message lifetime. `bus inbox` consumes `road-inbox`, not append-only history.
+
+### Configurable variables
+
+| Variable | Default | Use |
+|---|---|---|
+| `AGENTS_CITY_HOME` | `~/.agents-city` | root for all data and runtimes |
+| `AGENTS_CITY_USER` | resolved local identity | force an owner for tests/migrations |
+| `AGENTS_CITY_DATA` | selected city | force a city folder from an outside terminal |
+| `CITY_CODE_DIR` | `~/codigo` | destination for accepted GitHub clones |
+| `CITY_SEARCH_IN` | common home roots | colon-separated roots for local repo search |
+| `CITY_SEARCH_DEPTH` | `4` | maximum depth of that search |
+| `AGENTS_CITY_ORG` | empty | filter repositories by organisation; empty means all |
+| `CITY_SETTLE` | `8` | initial Claude startup wait in seconds |
+| `CITY_STAGGER` | `1` | extra separation per Claude window |
+| `CITY_BUS_URL` | empty | optional remote bus endpoint |
+| `CITY_BUS_TOKEN` | empty | credential for remote transport/map |
+| `AGENTS_CITY_URL` | `CITY_BUS_URL` | reporting/map endpoint when separate |
+| `CITY_DIR` | `~/.claude/channels/city-bus` | compatibility folder for `.env` and hooks |
+| `CITY_HOOKS` | `city` | `everywhere` runs the conscience hooks in every Claude session, not only city runtimes |
+
+Examples:
+
+```bash
+CITY_SEARCH_IN="$HOME/clients:$HOME/code" \
+CITY_SEARCH_DEPTH=6 \
+agents-city seat --repos
+
+AGENTS_CITY_HOME="$(mktemp -d)" \
+AGENTS_CITY_USER=tester \
+agents-city cities create lab
+
+CITY_SETTLE=0 CITY_STAGGER=0 agents-city seat --city product
+```
+
+The local repository index is cached for one day at
+`$XDG_CACHE_HOME/agents-city/repos.tsv` or `~/.cache/agents-city/repos.tsv`. The
+Hall offers a refresh. From a terminal, if you change `CITY_SEARCH_IN` while the
+cache is still valid, remove **only that index file** and repeat `seat --repos`.
+
+Transport setting precedence is:
+
+1. an environment variable already present;
+2. a recognised key in `~/.claude/channels/city-bus/.env`;
+3. for the token only on macOS, the Keychain service `city@agents-city`.
+
+The `.env` loader accepts known keys only and cannot redefine `PATH`. The session
+injects `CITY_ADDRESS`, `CITY_BUS_ACTOR`, `CITY_RUNTIME_KIND`, and
+`CITY_AGENT_ROLE` to authenticate each window; do not store them as global
+configuration.
+
+## Security and trust boundaries
+
+- The plugin's conscience stays inside the city: every hook checks for a city
+  identity (`CITY_BUS_ACTOR`) first and is silent in plain Claude sessions —
+  installing the plugin does not enrol every conversation on the machine.
+  `CITY_HOOKS=everywhere` is the explicit machine-wide opt-in.
+- Each city's hub binds a random port on `127.0.0.1`; it is not published to the
+  LAN.
+- Every actor has its own token and role. The seat is `chair`; each repo is
+  `member`.
+- Members receive no road credentials, cannot call `road.send`, and have no
+  member-to-member route.
+- Only a `seat -> seat` envelope addressed to a declared road may leave a city.
+- An invitation contains identity/address, never the remote token.
+- Protocol text fields are limited to 64,000 characters and IDs/paths are
+  normalised before use.
+- Known runtimes use native APIs. Only `terminal:<command>` permits the visible
+  tmux fallback.
+- `report` and `tokens` are dry-run by default; sending requires `--push`.
+- `reset` and `exit` have `--dry-run`; reset creates a backup, and neither should
+  touch unrelated tmux sessions.
+
+### The cage, the broker, and the audit chain
+
+Yolo mode stays — a committee cannot work if every bus command needs a human —
+but "do not ask" and "touch everything" are different axes, and only the first
+one is yolo. On macOS Claude, OpenCode and Kimi repo windows launch inside a
+generated seatbelt profile: writes land only in their own repo and runtime state, and the files
+that turn a prompt injection into a credential theft (`~/.ssh`,
+`~/.git-credentials`, `~/.aws`, gh and cloud configs, remote road tokens) are
+sealed at the kernel — reads and writes, children and grandchildren included.
+The agent is never asked anything; forbidden paths simply do not exist for it.
+Codex instead applies its native `workspace-write` sandbox and is not wrapped
+in seatbelt: MCP workers such as `node_repl` apply their own sandbox, and macOS
+rejects that operation inside an already-caged process. `CITY_CAGE=0`
+deliberately disables the applicable confinement layer.
+
+Because a caged window cannot read the `gh` token, PRs and pushes go through
+an opt-in credential broker (`CITY_BROKER=1`): a small owner-side process that
+holds the credentials, accepts per-window tokens bound to a single repo,
+refuses any action on the default branch, and writes every request — served or
+refused — to a hash-chained audit log the windows cannot touch. One rewritten
+byte breaks the chain and `broker.py verify` says so. The live kernel checks
+and both broker paths, happy and refused, run in `bin/test-cage.py` and
+`bin/test-broker.py`. The full model, its dials and its honest limits are in
+[docs/security.md](docs/security.md).
+
+Agents City isolates protocol responsibilities; it is not a hostile sandbox
+against the operating-system owner. Another process running as your user can
+read your repos, attach to tmux, or read private files in your home. Use separate
+accounts, VMs, or containers for untrusted code, and also apply each provider
+CLI's permission controls.
+
+A remote bus expands the trust surface. Deploy HTTPS/WSS, rotate tokens, limit
+scopes, and read [docs/self-host.md](docs/self-host.md). A road authorises message
+exchange between seats; it neither authorises execution of received commands nor
+grants remote filesystem access.
+
+## Troubleshooting
+
+### `agents-city seat` returns to an already open session
+
+This is expected. The tmux name is stable per owner/city. Detach with `Ctrl-b d`
+or inspect before closing:
+
+```bash
+agents-city exit <city> --dry-run
+```
+
+### I updated the package but still see the old behaviour
+
+A new global npm package does not replace live processes or tmux sessions. Check
+the binary being executed and restart only the city:
+
+```bash
+type -a agents-city
+agents-city --version
+npm root -g
+agents-city exit <city> --dry-run
+agents-city exit <city>
+agents-city seat --city <city>
+```
+
+With `fnm`, `nvm`, or `asdf`, every Node version may have a different global
+package set. Install the tarball under the Node version that will execute it.
+
+### Codex shows `city>` instead of its TUI
+
+Codex should show its official TUI. Verify a Codex version supporting
+`app-server`/`--remote`, update Agents City, and restart the city. Startup logs
+should show the WebSocket endpoint, the wait for the TUI thread, `Codex TUI
+thread ... adopted over WebSocket`, and bus authentication. After the first turn
+it will also show `joined over WebSocket`. `city>` is currently the expected
+console for OpenCode and Kimi.
+
+If you see `Failed to resume session ... no rollout found`, you are running the
+broken `0.3.0-beta.10` path, which tried to open a newly created thread through
+`codex resume --remote`. `0.3.0-beta.11` opened the right TUI but could wait
+indefinitely for an empty thread to materialize its first rollout. Install
+`0.3.0-beta.21` or later and restart only that
+city with `agents-city exit <city>` followed by `agents-city seat --city <city>`.
+
+### The `seat` window prints `fatal: not a git repository`
+
+The seat lives in the city data folder, which need not be a Git repository.
+Current launchers skip sync there. Seeing this error before Codex starts usually
+means the session still runs an old launcher: update, run `exit <city>`, and open
+it again. A real repo window without `.git` should be diagnosed separately.
+
+### Claude says the plugin is not on the Channels allowlist
+
+Agents City `0.3.0-beta.21` and later do not launch Claude with `--channels`.
+That message therefore identifies an old live session or a manual Channel
+invocation, not missing personal-account configuration. Do **not** create a
+machine-wide managed settings file or use `sudo`. Update Agents City, verify the
+version, then restart only the affected city with `agents-city exit <city>` and
+`agents-city seat --city <city>`. The normal log should say
+`Claude Code ready over persistent stream-json` and
+`claude-stream-json ready`.
+
+### Claude shows `Claude API` or asks for usage credits on a Team/Max account
+
+An inherited `CLAUDE_CODE_OAUTH_TOKEN`, API key, gateway URL, Bedrock, Vertex,
+or Foundry selector can take precedence over the healthy Claude.ai login stored
+by the CLI. Inspect names only — never print credential values:
+
+```bash
+claude auth status
+tmux show-environment -g | cut -d= -f1 | \
+  grep -E 'CLAUDE_CODE_OAUTH_TOKEN|ANTHROPIC_(API_KEY|AUTH_TOKEN|BASE_URL)|CLAUDE_CODE_USE_'
+env -u CLAUDE_CODE_OAUTH_TOKEN \
+    -u ANTHROPIC_API_KEY -u ANTHROPIC_AUTH_TOKEN -u ANTHROPIC_BASE_URL \
+    -u CLAUDE_CODE_USE_BEDROCK -u CLAUDE_CODE_USE_VERTEX \
+    -u CLAUDE_CODE_USE_FOUNDRY claude auth status
+```
+
+If the last command reports `authMethod: claude.ai`, Agents City uses that login
+and removes only those overrides from each new city child. It never deletes a
+token, logs out, or rewrites the credential store. Restart only that city after
+updating. To deliberately use environment/API authentication instead:
+
+```bash
+CITY_CLAUDE_AUTH=environment agents-city seat --city <city>
+```
+
+### Codex tools fail with `sandbox_apply: Operation not permitted`
+
+This is the macOS nested-sandbox failure: an older launcher put Codex's own
+sandbox (or an MCP worker sandbox) inside the city's seatbelt cage. Install the
+current beta and restart only the affected city. Codex now runs without the
+outer seatbelt and keeps its native `workspace-write` confinement, so repo reads
+and tools work without a nested `sandbox_apply`.
+
+### Codex reports an MCP executable is missing
+
+Codex inherits your global MCP registry. Agents City checks that registry
+without printing its environment values. If an enabled stdio MCP points to an
+executable that provably does not exist, it is disabled only for that city
+process; `~/.codex/config.toml` is not changed and healthy MCPs remain enabled.
+Inspect the scoped decision with:
+
+```bash
+agents-city logs --diagnostics | grep codex.mcp.unavailable.disabled
+```
+
+Fix or remove the original global entry later with `codex mcp`; a URL failure
+or another uncertain startup error is left visible rather than guessed away.
+
+### Text or JSON appears to be pasted into a window
+
+Claude, Codex, OpenCode, and Kimi do not use the clipboard or `tmux paste`.
+Inspect the card through:
+
+```bash
+agents-city seat --engines
+```
+
+If that row uses `terminal:<command>`, you selected the compatibility adapter and
+visible injection is expected. If a known runtime appears that way, select its
+native runtime again.
+
+### A Claude window shows a command ending in `--da`, `--dangerously`, or `-`
+
+That is a truncated legacy launch command, not a Claude or WebSocket message.
+Current versions write the full command into a private audited launcher and type
+only its short path into tmux. Update the package and restart only that city:
+
+```bash
+agents-city --version
+agents-city exit <city> --dry-run
+agents-city exit <city>
+agents-city seat --city <city>
+agents-city logs --diagnostics --follow
+```
+
+A failed launcher records `launch.failed`, prints its exit code and log path in
+the pane, and emits `runtime.launch.failed` to City live. It never logs the full
+command or credentials.
+
+### My repositories do not appear
+
+```bash
+command -v git
+git -C /path/to/repo remote get-url origin
+CITY_SEARCH_IN="/root/one:/root/two" \
+CITY_SEARCH_DEPTH=6 \
+agents-city seat --repos
+```
+
+Discovery requires `.git` (a directory or worktree file) and an `origin` remote.
+If roots just changed, refresh in the Hall or remove only
+`~/.cache/agents-city/repos.tsv`. `AGENTS_CITY_ORG` may be filtering the repo;
+leave it empty to index every remote.
+
+### GitHub does not show private repos or organisations
+
+```bash
+gh auth status
+gh api user --jq .login
+gh auth refresh -s read:org
+```
+
+Test access directly with `gh`; Agents City only consumes that session. An SSO
+organisation may require authorising the token in GitHub. Disk selection always
+works without OAuth.
+
+### A repo on the card gets no window
+
+The card may reference a repo that is not cloned. Run `seat --repos` and accept
+the clone, or clone it manually under an indexed root. If two names normalise to
+the same actor (for example, they differ only in punctuation), Agents City
+rejects the collision instead of mixing credentials.
+
+### The bus says it is already starting or running
+
+Do not delete locks while a process is alive. Inspect scope:
+
+```bash
+agents-city exit <city> --dry-run
+```
+
+If the expected session exists, reattach with `seat`. If it is a managed orphan,
+`exit <city>` stops it. The hub recovers a stale lock whose PID no longer exists;
+a newly created unreadable lock is preserved to prevent two simultaneous hubs.
+
+### An agent was offline when a task arrived
+
+Accepted internal envelopes remain in its outbox until ACK, with a 72-hour TTL.
+Reopen the same city/runtime to drain the queue. If the provider rejected the
+task, delivery remains failed and no ACK is invented. Use logs and the runtime
+benchmark to distinguish native rejection, authentication, and latency.
+
+### A local road exists but the destination is offline
+
+`road connect` configures reachability; it does not start the other city. Open
+both sessions:
+
+```bash
+agents-city seat --city source
+agents-city seat --city destination
+```
+
+A remote road additionally requires valid `CITY_BUS_URL` and `CITY_BUS_TOKEN` at
+both ends. A queued message does not mean the destination accepted or agreed.
+
+### The committee rejects my command
+
+Inspect state and schema first:
+
+```bash
+agents-city committee show <id>
+agents-city committee schema <verb>
+```
+
+Common rejections are deliberate: the seat tries to respond as a member; a
+member tries to decide; a position is missing without `--proceed-without`; a
+floor request remains pending; verifier equals executor when another identity is
+available; or closure is attempted before a passing verification.
+
+### I want to start over
+
+Do not delete all of `~/.agents-city` when only one city needs a reset:
+
+```bash
+agents-city reset <city> --dry-run
+agents-city reset <city>
+agents-city seat --city <city>
+```
+
+The output names the backup. Use `exit` when you only need to restart processes.
+
+## Development and testing
+
+### Full validation
+
+```bash
+git clone https://github.com/jlcases/agents-city.git
+cd agents-city
+npm install
+npm test
+```
+
+`npm test` runs `./bin/test`: Python/shell suites, buses and native runtimes with
+deterministic doubles, one throwaway tmux for unknown fallback only, a 40-actor
+stress test, cross-component contracts, and tarball allowlist checks. The default
+run is offline and uses temporary homes/repos.
+
+Focused tests:
+
+```bash
+./bin/test seat runtime-ui
+./bin/test channel committee runtime runtime-failures
+./bin/test stress benchmark contracts exit
+```
+
+### Typecheck and bundles
+
+```bash
+cd city/web
+npm run typecheck
+npm run build
+
+cd ../../plugin/channel
+npm run typecheck
+npm run build
+```
+
+Generated `plugin/channel` JavaScript ships in the package. Editing TypeScript
+without rebuilding leaves the tarball running old code.
+
+### Validate the exact package without publishing
+
+```bash
+npm pack --dry-run
+npm pack
+
+CITY_TEST_PREFIX="$(mktemp -d)"
+npm install -g --prefix "$CITY_TEST_PREFIX" ./agents-city-0.3.0-beta.21.tgz
+"$CITY_TEST_PREFIX/bin/agents-city" --version
+"$CITY_TEST_PREFIX/bin/agents-city" --help
+```
+
+For onboarding tests, also use temporary `HOME`, `AGENTS_CITY_HOME`, and
+`AGENTS_CITY_USER` values. Never point a suite at real city data.
+
+The full matrix and invariants live in [docs/testing.md](docs/testing.md).
+Benchmarks have dedicated guides in
+[benchmarks/stress/README.md](benchmarks/stress/README.md),
+[benchmarks/latency/README.md](benchmarks/latency/README.md), and
+[benchmarks/committee/README.md](benchmarks/committee/README.md).
+
+## Editions, license, and trust
+
+This repository is the **Community Edition**, licensed under
+[Apache-2.0](LICENSE): free to use, modify, self-host and build on, with an
+explicit patent grant. The license does not grant rights to the *Agents City*
+name — code travels, the name stays.
+
+An **Enterprise Edition** exists on top of this core: semantic city memory
+(vector search across acts, notices and deliberations), SSO, cross-city audit
+and fleet management. It is not in this repository. Agents City is built by
+[Arkatai](https://arkatai.com), an agentic-development studio — for the
+Enterprise Edition write to <hello@arkatai.com> or open an issue tagged
+`enterprise`.
+
+**No telemetry.** The product phones nobody home: everything runs on loopback
+and local files, and nothing about your work leaves the machine except what
+you configure yourself — a remote road, or a `--push` to your own worker. The
+only third-party request the web pages make is loading their typefaces from
+Google Fonts.