diff --git a/docs/tools/hivectl-cli.md b/docs/tools/hivectl-cli.md
index 507960f1..ac5cbc80 100644
--- a/docs/tools/hivectl-cli.md
+++ b/docs/tools/hivectl-cli.md
@@ -63,18 +63,18 @@ Sibling to the `hive-c0re` daemon binary. Covers host-side admin operations that
 
 ###### **Subcommands:**
 
-* `forge` — Forgejo user provisioning. Manual entry point to the same idempotent flow c0re runs automatically at boot (`forge::ensure_all`) — useful for recovery, ad-hoc reprovisioning, or single-agent fixes without bouncing the daemon
-* `matrix` — matrix-tuwunel user provisioning. Manual entry point to the same idempotent flow c0re runs automatically at boot (`matrix::ensure_all`) — useful when the boot-time sweep skipped an agent (e.g. matrix container wasn't up yet) or to re-register after wiping a token file
-* `github` — GitHub account provisioning: write an operator-supplied personal access token (PAT) into an agent's `github-token` state file so its `gh` wrapper + git credential helper can authenticate. Unlike forge/matrix there is no account creation — the operator supplies a PAT for an existing GitHub account. A CLI alternative to the dashboard credentials tab; the integration is on by default (`hyperhive.github.enable`), so no per-agent config is needed
-* `gateway` — Gateway htpasswd user management. Add, remove, or list users in an htpasswd file used by the gateway's HTTP Basic auth (`services.hyperhive.gateway.auth`). Credentials are stored as `BCrypt` hashes — no extra service or PAM required
-* `agents` — Agent container management. Requires the hive-c0re daemon to be running (connects to the host admin socket)
-* `approvals` — Operator approval queue: list pending requests, approve / deny them. Requires the hive-c0re daemon to be running (connects to the host admin socket)
-* `wg` — WireGuard inter-hive mesh setup helpers (`services.hyperhive.swarm`)
-* `peer-config` — Generate the federation peer-config block for THIS hive — the nix a peer operator pastes into their `services.hyperhive.swarm.peers` to trust + reach this hive. Emits `caCert` (+ a `cp` line for the cert) when this hive serves a self-signed CA, the WireGuard public key when the mesh key exists, and the `wireguard{Address,Endpoint}` you pass. The hive's own domain is filled in automatically from the running daemon (`services.hyperhive.domain`). Reads local state (the TLS CA cert, the wg key); never mutates. `wg init` calls this at the end, so a fresh mesh setup prints the hand-over block too
+* `forge` — Forgejo user provisioning
+* `matrix` — matrix-tuwunel user provisioning
+* `github` — GitHub account provisioning
+* `gateway` — Gateway htpasswd user management
+* `agents` — Agent container management
+* `approvals` — Operator approval queue: list, approve, or deny pending requests
+* `wg` — WireGuard inter-hive mesh setup helpers
+* `peer-config` — Generate the federation peer-config block for THIS hive
 * `choom` — Open an interactive Claude session inside an agent container
-* `stop` — Stop containers hive-wide in one operator action. Bare `hivectl stop` stops **everything** — all sub-agents plus the ci, forge, gateway, and matrix infra containers. Narrow it with scope flags: `--agents` (all sub-agents), `--ci` / `--forge` / `--gateway` / `--matrix` (named infra), and `--agent <name>` (repeatable) for specific sub-agents. Flags are additive (e.g. `--agents --matrix`). Requires the hive-c0re daemon (connects to the host admin socket). hive-c0re itself is never stopped — it services the request
-* `start` — Start containers hive-wide — the inverse of `hivectl stop`. Bare `hivectl start` restores the previously-running agents from the last broad-scope `stop`, or starts everything if no prior stop exists; the same scope flags as `stop` narrow it (`--agents`, `--ci`, `--forge`, `--gateway`, `--matrix`, `--agent <name>`). A targeted `--agent <name>` start is never filtered. Requires the hive-c0re daemon
-* `restart` — Restart containers hive-wide — `stop` then `start` over the same scope. Bare `hivectl restart` restarts **everything** (all sub-agents plus the ci/forge/gateway/matrix infra containers); the same scope flags as `stop`/`start` narrow it (`--agents`, `--ci`, `--forge`, `--gateway`, `--matrix`, `--agent <name>`). If the stop phase reports a failure the start phase is skipped so the operator can investigate. Requires the hive-c0re daemon
+* `stop` — Stop containers hive-wide in one operator action
+* `start` — Start containers hive-wide — the inverse of `hivectl stop`
+* `restart` — Restart containers hive-wide — `stop` then `start` over one scope
 * `quota` — Per-agent disk accounting + optional quotas via btrfs qgroups
 * `subvol` — btrfs subvolume management for agent state dirs
 * `open` — Print (and best-effort open in a browser) a hive web surface URL
@@ -90,7 +90,9 @@ Sibling to the `hive-c0re` daemon binary. Covers host-side admin operations that
 
 ## `hivectl forge`
 
-Forgejo user provisioning. Manual entry point to the same idempotent flow c0re runs automatically at boot (`forge::ensure_all`) — useful for recovery, ad-hoc reprovisioning, or single-agent fixes without bouncing the daemon
+Forgejo user provisioning.
+
+Manual entry point to the same idempotent provisioning c0re runs at boot — for recovery, ad-hoc reprovisioning, or fixing one agent without bouncing the daemon.
 
 **Usage:** `hivectl forge <COMMAND>`
 
@@ -104,11 +106,7 @@ Forgejo user provisioning. Manual entry point to the same idempotent flow c0re r
 
 Create or refresh the Forgejo account + token for `<name>`.
 
-When `<name>` matches an existing agent (i.e. it has a state dir under `/var/lib/hyperhive/agents/`), persists the token to `<state>/forge-token` (idempotent: re-mints + rewrites every call so the on-disk scope matches the current `forge::TOKEN_SCOPES`).
-
-When `<name>` is **not** an agent (a human or any other non-container account), creates the forgejo user and prints the freshly-minted token to stdout — no `/var/lib/hyperhive/agents/` directory is created for the user.
-
-Without `--password` / `--password-stdin` a random throwaway is used (fine for agents — they auth by token via tea / hive-forge). Set a password to log into the forge web UI afterwards. `--password` is idempotent: re-running with the same value sets the same password (covers password resets on already-created accounts since `forgejo admin user create` silently no-ops once the user exists).
+For an existing agent, persists the token to its state dir; for a human/other account, prints the token to stdout. Set a password to enable forge web-UI login (a random throwaway is used otherwise).
 
 **Usage:** `hivectl forge create-user [OPTIONS] <NAME>`
 
@@ -125,17 +123,19 @@ Without `--password` / `--password-stdin` a random throwaway is used (fine for a
 
 ## `hivectl matrix`
 
-matrix-tuwunel user provisioning. Manual entry point to the same idempotent flow c0re runs automatically at boot (`matrix::ensure_all`) — useful when the boot-time sweep skipped an agent (e.g. matrix container wasn't up yet) or to re-register after wiping a token file
+matrix-tuwunel user provisioning.
+
+Manual entry point to the same idempotent provisioning c0re runs at boot — for re-registering an agent the boot sweep skipped, or after wiping a token file.
 
 **Usage:** `hivectl matrix <COMMAND>`
 
 ###### **Subcommands:**
 
 * `create-user` — Create or refresh the matrix account + access token for `<name>`
-* `sync-admin` — Provision (or re-provision) the hive system admin matrix account (`@hive:<server>`). hive-c0re runs this automatically on startup before the agent sweep so the account is the first registered user — Conduit/tuwunel grants admin rights to the first user. Run manually to recover a missing admin token file
-* `promote-user` — Promote a matrix user to homeserver admin via the admin API. Uses the hive system admin token at `/var/lib/hyperhive/matrix/admin-token`. The `server_name` is discovered automatically from the running homeserver
-* `reset-password` — Reset a matrix user's password via the admin API and persist the new password to `/var/lib/hyperhive/matrix/creds/<name>-password` so the next `ensure_user_for` (or `create-user`) can re-login
-* `invite` — Invite a matrix user to the hive Space (default) or a specific room. Uses the hive admin token; the admin account must be a member of the target room with invite power (it owns the hive Space). Idempotent — already-member / already-invited is a no-op
+* `sync-admin` — Provision (or re-provision) the hive system admin matrix account
+* `promote-user` — Promote a matrix user to homeserver admin
+* `reset-password` — Reset a matrix user's password via the admin API
+* `invite` — Invite a matrix user to the hive Space, or a specific room with `--room`. Idempotent
 
 
 
@@ -143,11 +143,7 @@ matrix-tuwunel user provisioning. Manual entry point to the same idempotent flow
 
 Create or refresh the matrix account + access token for `<name>`.
 
-When `<name>` matches an existing agent (i.e. it has a state dir under `/var/lib/hyperhive/agents/`), persists the token to `<state>/matrix-token`. Skips registration when the file is already populated; delete it to force re-registration.
-
-When `<name>` is **not** an agent (a human or any other non-container account), registers the matrix user and prints the freshly-minted access token to stdout — no `/var/lib/hyperhive/agents/` directory is created for the user.
-
-Without `--password` / `--password-stdin` a random throwaway is used (fine for agents — they auth by `access_token`, never by password). Set a password to log into a matrix web client afterwards.
+For an existing agent, persists the token to its state dir; for a human/other account, prints the access token to stdout. Set a password to enable matrix web-client login (a random throwaway is used otherwise).
 
 **Usage:** `hivectl matrix create-user [OPTIONS] <NAME>`
 
@@ -164,7 +160,9 @@ Without `--password` / `--password-stdin` a random throwaway is used (fine for a
 
 ## `hivectl matrix sync-admin`
 
-Provision (or re-provision) the hive system admin matrix account (`@hive:<server>`). hive-c0re runs this automatically on startup before the agent sweep so the account is the first registered user — Conduit/tuwunel grants admin rights to the first user. Run manually to recover a missing admin token file
+Provision (or re-provision) the hive system admin matrix account.
+
+Runs automatically on startup; run manually to recover a missing admin token.
 
 **Usage:** `hivectl matrix sync-admin`
 
@@ -172,7 +170,7 @@ Provision (or re-provision) the hive system admin matrix account (`@hive:<server
 
 ## `hivectl matrix promote-user`
 
-Promote a matrix user to homeserver admin via the admin API. Uses the hive system admin token at `/var/lib/hyperhive/matrix/admin-token`. The `server_name` is discovered automatically from the running homeserver
+Promote a matrix user to homeserver admin
 
 **Usage:** `hivectl matrix promote-user <NAME>`
 
@@ -184,9 +182,9 @@ Promote a matrix user to homeserver admin via the admin API. Uses the hive syste
 
 ## `hivectl matrix reset-password`
 
-Reset a matrix user's password via the admin API and persist the new password to `/var/lib/hyperhive/matrix/creds/<name>-password` so the next `ensure_user_for` (or `create-user`) can re-login.
+Reset a matrix user's password via the admin API.
 
-After this command succeeds, run `hivectl matrix create-user <name>` to mint a fresh access token for the agent.
+Persists the new password so a later `create-user` can re-login.
 
 **Usage:** `hivectl matrix reset-password <NAME>`
 
@@ -198,7 +196,7 @@ After this command succeeds, run `hivectl matrix create-user <name>` to mint a f
 
 ## `hivectl matrix invite`
 
-Invite a matrix user to the hive Space (default) or a specific room. Uses the hive admin token; the admin account must be a member of the target room with invite power (it owns the hive Space). Idempotent — already-member / already-invited is a no-op
+Invite a matrix user to the hive Space, or a specific room with `--room`. Idempotent
 
 **Usage:** `hivectl matrix invite [OPTIONS] <USER>`
 
@@ -214,19 +212,23 @@ Invite a matrix user to the hive Space (default) or a specific room. Uses the hi
 
 ## `hivectl github`
 
-GitHub account provisioning: write an operator-supplied personal access token (PAT) into an agent's `github-token` state file so its `gh` wrapper + git credential helper can authenticate. Unlike forge/matrix there is no account creation — the operator supplies a PAT for an existing GitHub account. A CLI alternative to the dashboard credentials tab; the integration is on by default (`hyperhive.github.enable`), so no per-agent config is needed
+GitHub account provisioning.
+
+Store an operator-supplied personal access token (PAT) for an agent so its `gh` and git can authenticate. No account is created — the PAT is for an existing GitHub account.
 
 **Usage:** `hivectl github <COMMAND>`
 
 ###### **Subcommands:**
 
-* `set-token` — Write a GitHub PAT into `<agent>`'s state dir (`github-token`, 0600, agent-owned) via hive-priv. The GitHub integration is on by default (`hyperhive.github.enable`), so the `gh` wrapper + git credential helper pick the token up with no per-agent config. The token is read live at invocation, so no rebuild/restart is needed. Prefer `--token-stdin`: an inline `--token` is visible in shell history + process listings
+* `set-token` — Store a GitHub PAT for `<agent>` so its `gh` and git can authenticate
 
 
 
 ## `hivectl github set-token`
 
-Write a GitHub PAT into `<agent>`'s state dir (`github-token`, 0600, agent-owned) via hive-priv. The GitHub integration is on by default (`hyperhive.github.enable`), so the `gh` wrapper + git credential helper pick the token up with no per-agent config. The token is read live at invocation, so no rebuild/restart is needed. Prefer `--token-stdin`: an inline `--token` is visible in shell history + process listings
+Store a GitHub PAT for `<agent>` so its `gh` and git can authenticate.
+
+Prefer `--token-stdin` — an inline `--token` is visible in shell history.
 
 **Usage:** `hivectl github set-token [OPTIONS] <AGENT>`
 
@@ -243,23 +245,25 @@ Write a GitHub PAT into `<agent>`'s state dir (`github-token`, 0600, agent-owned
 
 ## `hivectl gateway`
 
-Gateway htpasswd user management. Add, remove, or list users in an htpasswd file used by the gateway's HTTP Basic auth (`services.hyperhive.gateway.auth`). Credentials are stored as `BCrypt` hashes — no extra service or PAM required
+Gateway htpasswd user management.
+
+Add, remove, or list users for the gateway's HTTP Basic auth.
 
 **Usage:** `hivectl gateway <COMMAND>`
 
 ###### **Subcommands:**
 
-* `create-user` — Add a new user or update the password of an existing user in the gateway htpasswd file. The daemon bcrypt-hashes the password (cost 12) and writes the credential store — hivectl relays over the host socket and never touches the file
-* `delete-user` — Remove a user from the gateway htpasswd file. Exits with an error when the user is not found so callers can detect the no-op case
+* `create-user` — Add a user or update an existing user's password in the gateway htpasswd
+* `delete-user` — Remove a user from the gateway htpasswd
 * `list-users` — List all gateway htpasswd usernames, one per line
 
 
 
 ## `hivectl gateway create-user`
 
-Add a new user or update the password of an existing user in the gateway htpasswd file. The daemon bcrypt-hashes the password (cost 12) and writes the credential store — hivectl relays over the host socket and never touches the file.
+Add a user or update an existing user's password in the gateway htpasswd.
 
-Pass `--password-stdin` when scripting or when you don't want the password visible in shell history.
+Use `--password-stdin` to keep the password out of shell history.
 
 **Usage:** `hivectl gateway create-user [OPTIONS] <USERNAME>`
 
@@ -276,7 +280,7 @@ Pass `--password-stdin` when scripting or when you don't want the password visib
 
 ## `hivectl gateway delete-user`
 
-Remove a user from the gateway htpasswd file. Exits with an error when the user is not found so callers can detect the no-op case
+Remove a user from the gateway htpasswd
 
 **Usage:** `hivectl gateway delete-user <USERNAME>`
 
@@ -296,27 +300,29 @@ List all gateway htpasswd usernames, one per line
 
 ## `hivectl agents`
 
-Agent container management. Requires the hive-c0re daemon to be running (connects to the host admin socket)
+Agent container management.
+
+Lifecycle actions on managed agent containers. Needs the hive-c0re daemon running.
 
 **Usage:** `hivectl agents <COMMAND>`
 
 ###### **Subcommands:**
 
-* `list` — Show all managed agents with their status (running / needs-login / needs-update) and technical state (deployed sha, parent, pending reminders). The host roster overview; reuses the dashboard's per-agent aggregation. Requires the daemon running
-* `restart` — Stop and start a single agent container without rebuilding config. Useful for "kick the container" when the process is stuck or the container needs a clean restart without changing the NixOS config. Rides the job queue (serialized against in-flight rebuilds for the same agent); waits with live progress unless `--no-wait`
-* `restart-all` — Restart ALL managed agent containers via one restart DAG each — unrelated agents overlap, each serializes on its own lease. Waits for the whole set with live progress unless `--no-wait`
-* `spawn` — Spawn a new agent container directly (`h-<name>`). Bypasses the approval queue — operator-on-the-host only. For approval-gated spawns, use `request-spawn`
-* `request-spawn` — Queue a spawn request as an approval. The container is created on `hivectl approvals approve <id>` (or the dashboard APPR0VE button)
+* `list` — Show all managed agents with their status and technical state
+* `restart` — Stop and start a single agent container without rebuilding config
+* `restart-all` — Restart all managed agent containers
+* `spawn` — Spawn a new agent container directly, bypassing the approval queue
+* `request-spawn` — Queue a spawn request for operator approval
 * `kill` — Stop a managed container (graceful)
-* `destroy` — Tear down a sub-agent container. The container is removed; persistent state (config repos + Claude credentials) is kept by default. Pass `--purge` to also wipe the agent's state dirs (config + creds + notes). No undo
+* `destroy` — Tear down a sub-agent container, keeping its state by default. No undo
 * `rebuild` — Apply pending config to a managed container
-* `set-parent` — Move an agent in the topology tree. Set `--parent` to a new parent agent name, or pass `--root` to promote the agent to root (no parent). Exactly one is required. Refuses cycles and unknown agents
+* `set-parent` — Move an agent in the topology tree — under a new parent, or to root
 
 
 
 ## `hivectl agents list`
 
-Show all managed agents with their status (running / needs-login / needs-update) and technical state (deployed sha, parent, pending reminders). The host roster overview; reuses the dashboard's per-agent aggregation. Requires the daemon running
+Show all managed agents with their status and technical state
 
 **Usage:** `hivectl agents list [OPTIONS]`
 
@@ -328,7 +334,7 @@ Show all managed agents with their status (running / needs-login / needs-update)
 
 ## `hivectl agents restart`
 
-Stop and start a single agent container without rebuilding config. Useful for "kick the container" when the process is stuck or the container needs a clean restart without changing the NixOS config. Rides the job queue (serialized against in-flight rebuilds for the same agent); waits with live progress unless `--no-wait`
+Stop and start a single agent container without rebuilding config
 
 **Usage:** `hivectl agents restart [OPTIONS] <NAME>`
 
@@ -344,7 +350,7 @@ Stop and start a single agent container without rebuilding config. Useful for "k
 
 ## `hivectl agents restart-all`
 
-Restart ALL managed agent containers via one restart DAG each — unrelated agents overlap, each serializes on its own lease. Waits for the whole set with live progress unless `--no-wait`
+Restart all managed agent containers
 
 **Usage:** `hivectl agents restart-all [OPTIONS]`
 
@@ -356,7 +362,9 @@ Restart ALL managed agent containers via one restart DAG each — unrelated agen
 
 ## `hivectl agents spawn`
 
-Spawn a new agent container directly (`h-<name>`). Bypasses the approval queue — operator-on-the-host only. For approval-gated spawns, use `request-spawn`
+Spawn a new agent container directly, bypassing the approval queue.
+
+Operator-on-the-host only; use `request-spawn` for an approval-gated spawn.
 
 **Usage:** `hivectl agents spawn <NAME>`
 
@@ -368,7 +376,7 @@ Spawn a new agent container directly (`h-<name>`). Bypasses the approval queue 
 
 ## `hivectl agents request-spawn`
 
-Queue a spawn request as an approval. The container is created on `hivectl approvals approve <id>` (or the dashboard APPR0VE button)
+Queue a spawn request for operator approval
 
 **Usage:** `hivectl agents request-spawn <NAME>`
 
@@ -392,7 +400,7 @@ Stop a managed container (graceful)
 
 ## `hivectl agents destroy`
 
-Tear down a sub-agent container. The container is removed; persistent state (config repos + Claude credentials) is kept by default. Pass `--purge` to also wipe the agent's state dirs (config + creds + notes). No undo
+Tear down a sub-agent container, keeping its state by default. No undo
 
 **Usage:** `hivectl agents destroy [OPTIONS] <NAME>`
 
@@ -402,7 +410,7 @@ Tear down a sub-agent container. The container is removed; persistent state (con
 
 ###### **Options:**
 
-* `--purge`
+* `--purge` — Also wipe the agent's state dirs (config + creds + notes)
 
 
 
@@ -420,7 +428,7 @@ Apply pending config to a managed container
 
 ## `hivectl agents set-parent`
 
-Move an agent in the topology tree. Set `--parent` to a new parent agent name, or pass `--root` to promote the agent to root (no parent). Exactly one is required. Refuses cycles and unknown agents
+Move an agent in the topology tree — under a new parent, or to root
 
 **Usage:** `hivectl agents set-parent [OPTIONS] <CHILD>`
 
@@ -437,7 +445,9 @@ Move an agent in the topology tree. Set `--parent` to a new parent agent name, o
 
 ## `hivectl approvals`
 
-Operator approval queue: list pending requests, approve / deny them. Requires the hive-c0re daemon to be running (connects to the host admin socket)
+Operator approval queue: list, approve, or deny pending requests.
+
+Needs the hive-c0re daemon running.
 
 **Usage:** `hivectl approvals <COMMAND>`
 
@@ -483,23 +493,23 @@ Deny a pending request by id
 
 ## `hivectl wg`
 
-WireGuard inter-hive mesh setup helpers (`services.hyperhive.swarm`).
+WireGuard inter-hive mesh setup helpers.
 
-One-time-setup convenience so nobody has to remember the `wg` dance: `wg init` generates + stores this hive's private key and prints the public key plus the nix snippet to enable the mesh; `wg peer` prints the snippet to add a remote hive; `wg status` wraps `wg show`. The verbs own the imperative state (the key file); the printed nix goes into the operator's host config (kept in git), so nothing here mutates declarative config behind the operator's back.
+Generate this hive's mesh key and print the nix to enable the mesh, add a peer, or inspect live interface state.
 
 **Usage:** `hivectl wg <COMMAND>`
 
 ###### **Subcommands:**
 
-* `init` — Generate (if absent) this hive's WireGuard private key, print its public key, and print the nix snippet to enable the mesh. Idempotent: an existing key is reused, never clobbered (clobbering would break a live mesh). Share the printed public key with peer hives
-* `peer` — Print the nix snippet to add a peer hive to the mesh. Pure output — paste it into this hive's config. Get `<pubkey>` from the peer's `hivectl wg init`
-* `status` — Show the live mesh interface state (`wg show wg-hive`). Requires the mesh to be enabled + up
+* `init` — Generate this hive's WireGuard key (if absent) and print its public key plus the nix to enable the mesh
+* `peer` — Print the nix to add a peer hive to the mesh
+* `status` — Show the live mesh interface state
 
 
 
 ## `hivectl wg init`
 
-Generate (if absent) this hive's WireGuard private key, print its public key, and print the nix snippet to enable the mesh. Idempotent: an existing key is reused, never clobbered (clobbering would break a live mesh). Share the printed public key with peer hives
+Generate this hive's WireGuard key (if absent) and print its public key plus the nix to enable the mesh
 
 **Usage:** `hivectl wg init [OPTIONS]`
 
@@ -511,7 +521,7 @@ Generate (if absent) this hive's WireGuard private key, print its public key, an
 
 ## `hivectl wg peer`
 
-Print the nix snippet to add a peer hive to the mesh. Pure output — paste it into this hive's config. Get `<pubkey>` from the peer's `hivectl wg init`
+Print the nix to add a peer hive to the mesh
 
 **Usage:** `hivectl wg peer [OPTIONS] --pubkey <PUBKEY> --address <ADDRESS> <DOMAIN>`
 
@@ -529,7 +539,7 @@ Print the nix snippet to add a peer hive to the mesh. Pure output — paste it i
 
 ## `hivectl wg status`
 
-Show the live mesh interface state (`wg show wg-hive`). Requires the mesh to be enabled + up
+Show the live mesh interface state
 
 **Usage:** `hivectl wg status`
 
@@ -537,7 +547,9 @@ Show the live mesh interface state (`wg show wg-hive`). Requires the mesh to be
 
 ## `hivectl peer-config`
 
-Generate the federation peer-config block for THIS hive — the nix a peer operator pastes into their `services.hyperhive.swarm.peers` to trust + reach this hive. Emits `caCert` (+ a `cp` line for the cert) when this hive serves a self-signed CA, the WireGuard public key when the mesh key exists, and the `wireguard{Address,Endpoint}` you pass. The hive's own domain is filled in automatically from the running daemon (`services.hyperhive.domain`). Reads local state (the TLS CA cert, the wg key); never mutates. `wg init` calls this at the end, so a fresh mesh setup prints the hand-over block too
+Generate the federation peer-config block for THIS hive.
+
+Prints the nix a peer operator pastes into their swarm config to trust and reach this hive.
 
 **Usage:** `hivectl peer-config [OPTIONS]`
 
@@ -552,7 +564,7 @@ Generate the federation peer-config block for THIS hive — the nix a peer opera
 
 Open an interactive Claude session inside an agent container.
 
-Runs claude as the agent user from its state dir with the harness's settings / MCP / system prompt. Bare `choom <name>` is a fresh session; `--resume <session-id>` rejoins a prior one. Never collides with the harness's live session. Requires root + a running container. See `docs/tools/hivectl.md` (Choom) for details.
+A fresh session by default, or resume a prior one. Requires root and a running container.
 
 **Usage:** `hivectl choom [OPTIONS] <NAME>`
 
@@ -568,7 +580,9 @@ Runs claude as the agent user from its state dir with the harness's settings / M
 
 ## `hivectl stop`
 
-Stop containers hive-wide in one operator action. Bare `hivectl stop` stops **everything** — all sub-agents plus the ci, forge, gateway, and matrix infra containers. Narrow it with scope flags: `--agents` (all sub-agents), `--ci` / `--forge` / `--gateway` / `--matrix` (named infra), and `--agent <name>` (repeatable) for specific sub-agents. Flags are additive (e.g. `--agents --matrix`). Requires the hive-c0re daemon (connects to the host admin socket). hive-c0re itself is never stopped — it services the request
+Stop containers hive-wide in one operator action.
+
+Bare `hivectl stop` stops everything; scope flags narrow it to specific sub-agents or infra containers.
 
 **Usage:** `hivectl stop [OPTIONS]`
 
@@ -587,7 +601,9 @@ Stop containers hive-wide in one operator action. Bare `hivectl stop` stops **ev
 
 ## `hivectl start`
 
-Start containers hive-wide — the inverse of `hivectl stop`. Bare `hivectl start` restores the previously-running agents from the last broad-scope `stop`, or starts everything if no prior stop exists; the same scope flags as `stop` narrow it (`--agents`, `--ci`, `--forge`, `--gateway`, `--matrix`, `--agent <name>`). A targeted `--agent <name>` start is never filtered. Requires the hive-c0re daemon
+Start containers hive-wide — the inverse of `hivectl stop`.
+
+Bare `hivectl start` restores the agents stopped by the last broad-scope `stop` (or starts everything if none); scope flags narrow it.
 
 **Usage:** `hivectl start [OPTIONS]`
 
@@ -605,7 +621,9 @@ Start containers hive-wide — the inverse of `hivectl stop`. Bare `hivectl star
 
 ## `hivectl restart`
 
-Restart containers hive-wide — `stop` then `start` over the same scope. Bare `hivectl restart` restarts **everything** (all sub-agents plus the ci/forge/gateway/matrix infra containers); the same scope flags as `stop`/`start` narrow it (`--agents`, `--ci`, `--forge`, `--gateway`, `--matrix`, `--agent <name>`). If the stop phase reports a failure the start phase is skipped so the operator can investigate. Requires the hive-c0re daemon
+Restart containers hive-wide — `stop` then `start` over one scope.
+
+Bare `hivectl restart` restarts everything; scope flags narrow it.
 
 **Usage:** `hivectl restart [OPTIONS]`
 
@@ -625,21 +643,23 @@ Restart containers hive-wide — `stop` then `start` over the same scope. Bare `
 
 Per-agent disk accounting + optional quotas via btrfs qgroups.
 
-Opt-in: `quota enable` turns on btrfs qgroup accounting for the agent-state filesystem (a one-time, I/O-heavy rescan — that's why it isn't automatic). Then `quota show` reports per-agent usage and `quota limit` caps an agent. No-op on non-btrfs hosts. Operates on agent state subvolumes (created by the btrfs-subvolume migration); agents still on a plain dir report no qgroup usage.
+Opt-in: enable qgroup accounting, then report per-agent usage or cap an agent. No-op on non-btrfs hosts.
 
 **Usage:** `hivectl quota <COMMAND>`
 
 ###### **Subcommands:**
 
-* `enable` — Enable btrfs qgroup accounting on the agent-state filesystem. Run once before `show` / `limit`. Triggers a full btrfs rescan (I/O heavy on a large filesystem), so it's a deliberate opt-in. Idempotent; a no-op on non-btrfs hosts
-* `show` — Report per-agent disk usage (referenced + exclusive bytes) from btrfs qgroups. With no name, shows every agent that has a state subvolume; pass a name to show just that one. Requires `enable` first
-* `limit` — Set or clear an agent's disk quota (a referenced-usage cap). `size` accepts a byte count or a `K`/`M`/`G`/`T` suffix (e.g. `5G`), or `none` to clear the limit. Requires `enable` first
+* `enable` — Enable btrfs qgroup accounting on the agent-state filesystem
+* `show` — Report per-agent disk usage from btrfs qgroups (all agents, or one by name)
+* `limit` — Set or clear an agent's disk-usage quota
 
 
 
 ## `hivectl quota enable`
 
-Enable btrfs qgroup accounting on the agent-state filesystem. Run once before `show` / `limit`. Triggers a full btrfs rescan (I/O heavy on a large filesystem), so it's a deliberate opt-in. Idempotent; a no-op on non-btrfs hosts
+Enable btrfs qgroup accounting on the agent-state filesystem.
+
+Run once before `show` / `limit`. No-op on non-btrfs hosts.
 
 **Usage:** `hivectl quota enable`
 
@@ -647,7 +667,7 @@ Enable btrfs qgroup accounting on the agent-state filesystem. Run once before `s
 
 ## `hivectl quota show`
 
-Report per-agent disk usage (referenced + exclusive bytes) from btrfs qgroups. With no name, shows every agent that has a state subvolume; pass a name to show just that one. Requires `enable` first
+Report per-agent disk usage from btrfs qgroups (all agents, or one by name)
 
 **Usage:** `hivectl quota show [NAME]`
 
@@ -659,7 +679,7 @@ Report per-agent disk usage (referenced + exclusive bytes) from btrfs qgroups. W
 
 ## `hivectl quota limit`
 
-Set or clear an agent's disk quota (a referenced-usage cap). `size` accepts a byte count or a `K`/`M`/`G`/`T` suffix (e.g. `5G`), or `none` to clear the limit. Requires `enable` first
+Set or clear an agent's disk-usage quota
 
 **Usage:** `hivectl quota limit <NAME> <SIZE>`
 
@@ -674,20 +694,22 @@ Set or clear an agent's disk quota (a referenced-usage cap). `size` accepts a by
 
 btrfs subvolume management for agent state dirs.
 
-New agents get a btrfs subvolume state root automatically (when the host FS is btrfs); agents that predate that are left on plain dirs. `subvol upgrade <agent>` opts an existing plain-dir agent into the subvolume feature set (snapshots, per-subvol usage/quota, migration) by migrating its state dir in place. Requires the hive-c0re daemon (for the stop/start) and root (for the privileged migration).
+Upgrade an existing plain-dir agent's state into a btrfs subvolume so it gains snapshots and per-subvol usage/quota.
 
 **Usage:** `hivectl subvol <COMMAND>`
 
 ###### **Subcommands:**
 
-* `upgrade` — Convert an existing plain-dir agent state root into a btrfs subvolume in place. Stops the agent (so its state bind-mount is released), migrates `…/agents/<name>/` to a subvolume preserving ownership/permissions/xattrs, then restarts it. Idempotent (no-op if already a subvolume) and safe (the original dir is left untouched on any failure before the final swap). Requires `--yes` since it bounces the agent and moves its state
-* `snapshot` — Read-only snapshots of an agent's state subvolume — the first step of the (in-progress) inter-hive migration path, or a manual point-in-time backup
+* `upgrade` — Convert a plain-dir agent state root into a btrfs subvolume in place, so it gains snapshots and per-subvol usage/quota
+* `snapshot` — Read-only snapshots of an agent's state subvolume
 
 
 
 ## `hivectl subvol upgrade`
 
-Convert an existing plain-dir agent state root into a btrfs subvolume in place. Stops the agent (so its state bind-mount is released), migrates `…/agents/<name>/` to a subvolume preserving ownership/permissions/xattrs, then restarts it. Idempotent (no-op if already a subvolume) and safe (the original dir is left untouched on any failure before the final swap). Requires `--yes` since it bounces the agent and moves its state
+Convert a plain-dir agent state root into a btrfs subvolume in place, so it gains snapshots and per-subvol usage/quota.
+
+Bounces the agent to migrate its state, so it requires `--yes`.
 
 **Usage:** `hivectl subvol upgrade [OPTIONS] <NAME>`
 
@@ -703,13 +725,13 @@ Convert an existing plain-dir agent state root into a btrfs subvolume in place.
 
 ## `hivectl subvol snapshot`
 
-Read-only snapshots of an agent's state subvolume — the first step of the (in-progress) inter-hive migration path, or a manual point-in-time backup
+Read-only snapshots of an agent's state subvolume
 
 **Usage:** `hivectl subvol snapshot <COMMAND>`
 
 ###### **Subcommands:**
 
-* `create` — Create a read-only snapshot. Agent must already be a subvolume (`subvol upgrade` first). Prints the snapshot's host path
+* `create` — Create a read-only snapshot (agent must already be a subvolume)
 * `delete` — Delete a snapshot created by `subvol snapshot create`
 * `send` — Export a snapshot to a local file via `btrfs send` (the local-file half of inter-hive migration transport; the cross-hive `ssh ... btrfs receive` leg isn't wired up yet). Also useful standalone as a point-in-time backup: a full send with no `--parent` produces a self-contained archive of the snapshot
 
@@ -717,7 +739,7 @@ Read-only snapshots of an agent's state subvolume — the first step of the (in-
 
 ## `hivectl subvol snapshot create`
 
-Create a read-only snapshot. Agent must already be a subvolume (`subvol upgrade` first). Prints the snapshot's host path
+Create a read-only snapshot (agent must already be a subvolume)
 
 **Usage:** `hivectl subvol snapshot create --label <LABEL> <NAME>`
 
@@ -766,7 +788,7 @@ Export a snapshot to a local file via `btrfs send` (the local-file half of inter
 
 Print (and best-effort open in a browser) a hive web surface URL.
 
-Resolves the URL from the running daemon (`HostRequest::Urls`), so custom forge / matrix domains work without guessing `forge.<domain>`. Prints the URL unconditionally — the reliable core, since the host is usually headless / driven over SSH where `xdg-open` is a no-op — then tries `xdg-open` as a convenience. Bare `hivectl open` opens the operator dashboard.
+Resolves the URL from the running daemon so custom forge / matrix domains work. Bare `hivectl open` opens the operator dashboard.
 
 **Usage:** `hivectl open [TARGET]`
 
@@ -791,7 +813,7 @@ Resolves the URL from the running daemon (`HostRequest::Urls`), so custom forge
 
 Generate a shell completion script for `hivectl` and print it to stdout.
 
-Pipe it into your shell's completion path — e.g. for zsh: `hivectl completions zsh > ~/.zsh/completions/_hivectl` (with that dir on `$fpath`). The hyperhive NixOS module installs the zsh script system-wide automatically, so this is mainly for ad-hoc / other-shell use. Supports bash, zsh, fish, elvish, and powershell.
+Supports bash, zsh, fish, elvish, and powershell. The NixOS module already installs the zsh script system-wide; this is for ad-hoc or other-shell use.
 
 **Usage:** `hivectl completions <SHELL>`
 
diff --git a/hivectl/src/main.rs b/hivectl/src/main.rs
index 9f661739..74ed07bb 100644
--- a/hivectl/src/main.rs
+++ b/hivectl/src/main.rs
@@ -53,77 +53,67 @@ struct Cli {
 
 #[derive(Subcommand)]
 enum Cmd {
-    /// Forgejo user provisioning. Manual entry point to the same
-    /// idempotent flow c0re runs automatically at boot
-    /// (`forge::ensure_all`) — useful for recovery, ad-hoc reprovisioning,
-    /// or single-agent fixes without bouncing the daemon.
+    /// Forgejo user provisioning.
+    ///
+    /// Manual entry point to the same idempotent provisioning c0re runs at
+    /// boot — for recovery, ad-hoc reprovisioning, or fixing one agent
+    /// without bouncing the daemon.
     Forge {
         #[command(subcommand)]
         cmd: ForgeCmd,
     },
-    /// matrix-tuwunel user provisioning. Manual entry point to the same
-    /// idempotent flow c0re runs automatically at boot
-    /// (`matrix::ensure_all`) — useful when the boot-time sweep skipped
-    /// an agent (e.g. matrix container wasn't up yet) or to re-register
-    /// after wiping a token file.
+    /// matrix-tuwunel user provisioning.
+    ///
+    /// Manual entry point to the same idempotent provisioning c0re runs at
+    /// boot — for re-registering an agent the boot sweep skipped, or after
+    /// wiping a token file.
     Matrix {
         #[command(subcommand)]
         cmd: MatrixCmd,
     },
-    /// GitHub account provisioning: write an operator-supplied personal
-    /// access token (PAT) into an agent's `github-token` state file so its
-    /// `gh` wrapper + git credential helper can authenticate. Unlike
-    /// forge/matrix there is no account creation — the operator supplies a
-    /// PAT for an existing GitHub account. A CLI alternative to the dashboard
-    /// credentials tab; the integration is on by default
-    /// (`hyperhive.github.enable`), so no per-agent config is needed.
+    /// GitHub account provisioning.
+    ///
+    /// Store an operator-supplied personal access token (PAT) for an agent
+    /// so its `gh` and git can authenticate. No account is created — the
+    /// PAT is for an existing GitHub account.
     Github {
         #[command(subcommand)]
         cmd: GithubCmd,
     },
-    /// Gateway htpasswd user management. Add, remove, or list users in
-    /// an htpasswd file used by the gateway's HTTP Basic auth
-    /// (`services.hyperhive.gateway.auth`). Credentials are stored as
-    /// `BCrypt` hashes — no extra service or PAM required.
+    /// Gateway htpasswd user management.
+    ///
+    /// Add, remove, or list users for the gateway's HTTP Basic auth.
     Gateway {
         #[command(subcommand)]
         cmd: GatewayCmd,
     },
-    /// Agent container management. Requires the hive-c0re daemon to be
-    /// running (connects to the host admin socket).
+    /// Agent container management.
+    ///
+    /// Lifecycle actions on managed agent containers. Needs the hive-c0re
+    /// daemon running.
     Agents {
         #[command(subcommand)]
         cmd: AgentsCmd,
     },
-    /// Operator approval queue: list pending requests, approve / deny them.
-    /// Requires the hive-c0re daemon to be running (connects to the host
-    /// admin socket).
+    /// Operator approval queue: list, approve, or deny pending requests.
+    ///
+    /// Needs the hive-c0re daemon running.
     Approvals {
         #[command(subcommand)]
         cmd: ApprovalsCmd,
     },
-    /// WireGuard inter-hive mesh setup helpers (`services.hyperhive.swarm`).
+    /// WireGuard inter-hive mesh setup helpers.
     ///
-    /// One-time-setup convenience so nobody has to remember the `wg` dance:
-    /// `wg init` generates + stores this hive's private key and prints the
-    /// public key plus the nix snippet to enable the mesh; `wg peer` prints
-    /// the snippet to add a remote hive; `wg status` wraps `wg show`. The
-    /// verbs own the imperative state (the key file); the printed nix goes
-    /// into the operator's host config (kept in git), so nothing here mutates
-    /// declarative config behind the operator's back.
+    /// Generate this hive's mesh key and print the nix to enable the mesh,
+    /// add a peer, or inspect live interface state.
     Wg {
         #[command(subcommand)]
         cmd: WgCmd,
     },
-    /// Generate the federation peer-config block for THIS hive — the nix
-    /// a peer operator pastes into their `services.hyperhive.swarm.peers`
-    /// to trust + reach this hive. Emits `caCert` (+ a `cp` line for the
-    /// cert) when this hive serves a self-signed CA, the WireGuard public
-    /// key when the mesh key exists, and the `wireguard{Address,Endpoint}`
-    /// you pass. The hive's own domain is filled in automatically from the
-    /// running daemon (`services.hyperhive.domain`). Reads local state (the
-    /// TLS CA cert, the wg key); never mutates. `wg init` calls this at the
-    /// end, so a fresh mesh setup prints the hand-over block too.
+    /// Generate the federation peer-config block for THIS hive.
+    ///
+    /// Prints the nix a peer operator pastes into their swarm config to
+    /// trust and reach this hive.
     PeerConfig {
         /// This hive's WireGuard mesh address (e.g. `10.42.0.1/32`),
         /// emitted as `wireguardAddress`. Omit when not running the mesh.
@@ -136,11 +126,8 @@ enum Cmd {
     },
     /// Open an interactive Claude session inside an agent container.
     ///
-    /// Runs claude as the agent user from its state dir with the harness's
-    /// settings / MCP / system prompt. Bare `choom <name>` is a fresh
-    /// session; `--resume <session-id>` rejoins a prior one. Never
-    /// collides with the harness's live session. Requires root + a running
-    /// container. See `docs/tools/hivectl.md` (Choom) for details.
+    /// A fresh session by default, or resume a prior one. Requires root
+    /// and a running container.
     Choom {
         /// Agent name (e.g. `damocles`, `iris`).
         name: String,
@@ -153,14 +140,10 @@ enum Cmd {
         #[arg(long = "resume", value_name = "SESSION")]
         resume_session: Option<String>,
     },
-    /// Stop containers hive-wide in one operator action. Bare `hivectl
-    /// stop` stops **everything** — all sub-agents plus the ci, forge,
-    /// gateway, and matrix infra containers. Narrow it with scope flags:
-    /// `--agents` (all sub-agents), `--ci` / `--forge` / `--gateway` /
-    /// `--matrix` (named infra), and `--agent <name>` (repeatable) for
-    /// specific sub-agents. Flags are additive (e.g. `--agents --matrix`).
-    /// Requires the hive-c0re daemon (connects to the host admin socket).
-    /// hive-c0re itself is never stopped — it services the request.
+    /// Stop containers hive-wide in one operator action.
+    ///
+    /// Bare `hivectl stop` stops everything; scope flags narrow it to
+    /// specific sub-agents or infra containers.
     Stop {
         #[command(flatten)]
         scope: ScopeArgs,
@@ -177,13 +160,11 @@ enum Cmd {
         #[arg(long)]
         no_wait: bool,
     },
-    /// Start containers hive-wide — the inverse of `hivectl stop`. Bare
-    /// `hivectl start` restores the previously-running agents from the
-    /// last broad-scope `stop`, or starts everything if no prior stop
-    /// exists; the same scope flags as `stop` narrow it (`--agents`,
-    /// `--ci`, `--forge`, `--gateway`, `--matrix`, `--agent <name>`). A
-    /// targeted `--agent <name>` start is never filtered. Requires the
-    /// hive-c0re daemon.
+    /// Start containers hive-wide — the inverse of `hivectl stop`.
+    ///
+    /// Bare `hivectl start` restores the agents stopped by the last
+    /// broad-scope `stop` (or starts everything if none); scope flags
+    /// narrow it.
     Start {
         #[command(flatten)]
         scope: ScopeArgs,
@@ -192,13 +173,9 @@ enum Cmd {
         #[arg(long)]
         no_wait: bool,
     },
-    /// Restart containers hive-wide — `stop` then `start` over the same
-    /// scope. Bare `hivectl restart` restarts **everything** (all sub-agents
-    /// plus the ci/forge/gateway/matrix infra containers); the same scope
-    /// flags as `stop`/`start` narrow it (`--agents`, `--ci`, `--forge`,
-    /// `--gateway`, `--matrix`, `--agent <name>`). If the stop phase reports
-    /// a failure the start phase is skipped so the operator can investigate.
-    /// Requires the hive-c0re daemon.
+    /// Restart containers hive-wide — `stop` then `start` over one scope.
+    ///
+    /// Bare `hivectl restart` restarts everything; scope flags narrow it.
     Restart {
         #[command(flatten)]
         scope: ScopeArgs,
@@ -209,36 +186,24 @@ enum Cmd {
     },
     /// Per-agent disk accounting + optional quotas via btrfs qgroups.
     ///
-    /// Opt-in: `quota enable` turns on btrfs qgroup accounting for the
-    /// agent-state filesystem (a one-time, I/O-heavy rescan — that's why
-    /// it isn't automatic). Then `quota show` reports per-agent usage and
-    /// `quota limit` caps an agent. No-op on non-btrfs hosts. Operates on
-    /// agent state subvolumes (created by the btrfs-subvolume migration);
-    /// agents still on a plain dir report no qgroup usage.
+    /// Opt-in: enable qgroup accounting, then report per-agent usage or
+    /// cap an agent. No-op on non-btrfs hosts.
     Quota {
         #[command(subcommand)]
         cmd: QuotaCmd,
     },
     /// btrfs subvolume management for agent state dirs.
     ///
-    /// New agents get a btrfs subvolume state root automatically (when the
-    /// host FS is btrfs); agents that predate that are left on plain dirs.
-    /// `subvol upgrade <agent>` opts an existing plain-dir agent into the
-    /// subvolume feature set (snapshots, per-subvol usage/quota, migration)
-    /// by migrating its state dir in place. Requires the hive-c0re daemon
-    /// (for the stop/start) and root (for the privileged migration).
+    /// Upgrade an existing plain-dir agent's state into a btrfs subvolume
+    /// so it gains snapshots and per-subvol usage/quota.
     Subvol {
         #[command(subcommand)]
         cmd: SubvolCmd,
     },
     /// Print (and best-effort open in a browser) a hive web surface URL.
     ///
-    /// Resolves the URL from the running daemon (`HostRequest::Urls`), so
-    /// custom forge / matrix domains work without guessing `forge.<domain>`.
-    /// Prints the URL unconditionally — the reliable core, since the host
-    /// is usually headless / driven over SSH where `xdg-open` is a no-op —
-    /// then tries `xdg-open` as a convenience. Bare `hivectl open` opens the
-    /// operator dashboard.
+    /// Resolves the URL from the running daemon so custom forge / matrix
+    /// domains work. Bare `hivectl open` opens the operator dashboard.
     Open {
         /// Which surface to open. Defaults to the operator dashboard.
         #[arg(value_enum, default_value_t = OpenTarget::Home)]
@@ -246,21 +211,16 @@ enum Cmd {
     },
     /// Emit the full CLI reference as `CommonMark` to stdout.
     ///
-    /// Hidden tooling command (not part of day-to-day operator admin):
-    /// walks this binary's own clap command tree and renders every verb,
-    /// flag, and help string as markdown. Used by the docs build to keep
-    /// the published `hivectl` reference in lockstep with the code — no
-    /// hand-maintained command list to drift out of date.
+    /// Hidden tooling command used by the docs build to keep the published
+    /// `hivectl` reference in lockstep with the code.
     #[command(hide = true)]
     MarkdownDocs,
     /// Generate a shell completion script for `hivectl` and print it to
     /// stdout.
     ///
-    /// Pipe it into your shell's completion path — e.g. for zsh:
-    /// `hivectl completions zsh > ~/.zsh/completions/_hivectl` (with that
-    /// dir on `$fpath`). The hyperhive NixOS module installs the zsh script
-    /// system-wide automatically, so this is mainly for ad-hoc / other-shell
-    /// use. Supports bash, zsh, fish, elvish, and powershell.
+    /// Supports bash, zsh, fish, elvish, and powershell. The NixOS module
+    /// already installs the zsh script system-wide; this is for ad-hoc or
+    /// other-shell use.
     Completions {
         /// Shell to emit completions for.
         shell: clap_complete::Shell,
@@ -325,24 +285,9 @@ impl ScopeArgs {
 enum ForgeCmd {
     /// Create or refresh the Forgejo account + token for `<name>`.
     ///
-    /// When `<name>` matches an existing agent (i.e. it has a state
-    /// dir under `/var/lib/hyperhive/agents/`), persists the token to
-    /// `<state>/forge-token` (idempotent: re-mints + rewrites every
-    /// call so the on-disk scope matches the current
-    /// `forge::TOKEN_SCOPES`).
-    ///
-    /// When `<name>` is **not** an agent (a human or any other
-    /// non-container account), creates the forgejo user and prints the
-    /// freshly-minted token to stdout — no `/var/lib/hyperhive/agents/`
-    /// directory is created for the user.
-    ///
-    /// Without `--password` / `--password-stdin` a random throwaway is
-    /// used (fine for agents — they auth by token via tea / hive-forge).
-    /// Set a password to log into the forge web UI afterwards.
-    /// `--password` is idempotent: re-running with the same value sets
-    /// the same password (covers password resets on already-created
-    /// accounts since `forgejo admin user create` silently no-ops once
-    /// the user exists).
+    /// For an existing agent, persists the token to its state dir; for a
+    /// human/other account, prints the token to stdout. Set a password to
+    /// enable forge web-UI login (a random throwaway is used otherwise).
     CreateUser {
         /// Forgejo username. For agents: the container/agent name
         /// (`<n>` in `h-<n>`; manager uses the literal `manager`).
@@ -368,20 +313,10 @@ enum ForgeCmd {
 enum MatrixCmd {
     /// Create or refresh the matrix account + access token for `<name>`.
     ///
-    /// When `<name>` matches an existing agent (i.e. it has a state
-    /// dir under `/var/lib/hyperhive/agents/`), persists the token to
-    /// `<state>/matrix-token`. Skips registration when the file is
-    /// already populated; delete it to force re-registration.
-    ///
-    /// When `<name>` is **not** an agent (a human or any other
-    /// non-container account), registers the matrix user and prints
-    /// the freshly-minted access token to stdout — no
-    /// `/var/lib/hyperhive/agents/` directory is created for the user.
-    ///
-    /// Without `--password` / `--password-stdin` a random throwaway is
-    /// used (fine for agents — they auth by `access_token`, never by
-    /// password). Set a password to log into a matrix web client
-    /// afterwards.
+    /// For an existing agent, persists the token to its state dir; for a
+    /// human/other account, prints the access token to stdout. Set a
+    /// password to enable matrix web-client login (a random throwaway is
+    /// used otherwise).
     CreateUser {
         /// Matrix localpart. For agents: the container/agent name.
         /// For humans: any matrix localpart — `mara`, `damocles`, etc.
@@ -400,34 +335,25 @@ enum MatrixCmd {
         #[arg(long, conflicts_with = "password")]
         password_stdin: bool,
     },
-    /// Provision (or re-provision) the hive system admin matrix account
-    /// (`@hive:<server>`). hive-c0re runs this automatically on startup
-    /// before the agent sweep so the account is the first registered
-    /// user — Conduit/tuwunel grants admin rights to the first user.
-    /// Run manually to recover a missing admin token file.
+    /// Provision (or re-provision) the hive system admin matrix account.
+    ///
+    /// Runs automatically on startup; run manually to recover a missing
+    /// admin token.
     SyncAdmin,
-    /// Promote a matrix user to homeserver admin via the admin API.
-    /// Uses the hive system admin token at
-    /// `/var/lib/hyperhive/matrix/admin-token`. The `server_name` is
-    /// discovered automatically from the running homeserver.
+    /// Promote a matrix user to homeserver admin.
     PromoteUser {
         /// Matrix localpart of the user to promote (e.g. `argus`).
         name: String,
     },
-    /// Reset a matrix user's password via the admin API and persist the
-    /// new password to `/var/lib/hyperhive/matrix/creds/<name>-password`
-    /// so the next `ensure_user_for` (or `create-user`) can re-login.
+    /// Reset a matrix user's password via the admin API.
     ///
-    /// After this command succeeds, run `hivectl matrix create-user
-    /// <name>` to mint a fresh access token for the agent.
+    /// Persists the new password so a later `create-user` can re-login.
     ResetPassword {
         /// Matrix localpart of the account to reset (e.g. `argus`).
         name: String,
     },
-    /// Invite a matrix user to the hive Space (default) or a specific
-    /// room. Uses the hive admin token; the admin account must be a
-    /// member of the target room with invite power (it owns the hive
-    /// Space). Idempotent — already-member / already-invited is a no-op.
+    /// Invite a matrix user to the hive Space, or a specific room with
+    /// `--room`. Idempotent.
     Invite {
         /// User to invite: a full id (`@mara:server`) or a bare
         /// localpart (qualified with the homeserver's `server_name`).
@@ -441,13 +367,11 @@ enum MatrixCmd {
 
 #[derive(Subcommand)]
 enum GithubCmd {
-    /// Write a GitHub PAT into `<agent>`'s state dir (`github-token`, 0600,
-    /// agent-owned) via hive-priv. The GitHub integration is on by default
-    /// (`hyperhive.github.enable`), so the `gh` wrapper + git credential
-    /// helper pick the token up with no per-agent config. The token is read
-    /// live at invocation, so no rebuild/restart is needed. Prefer
-    /// `--token-stdin`: an inline `--token` is visible in shell history +
-    /// process listings.
+    /// Store a GitHub PAT for `<agent>` so its `gh` and git can
+    /// authenticate.
+    ///
+    /// Prefer `--token-stdin` — an inline `--token` is visible in shell
+    /// history.
     SetToken {
         /// Logical agent name (the container/agent name).
         agent: String,
@@ -463,13 +387,10 @@ enum GithubCmd {
 
 #[derive(Subcommand)]
 enum GatewayCmd {
-    /// Add a new user or update the password of an existing user in the
-    /// gateway htpasswd file. The daemon bcrypt-hashes the password (cost 12)
-    /// and writes the credential store — hivectl relays over the host socket
-    /// and never touches the file.
+    /// Add a user or update an existing user's password in the gateway
+    /// htpasswd.
     ///
-    /// Pass `--password-stdin` when scripting or when you don't want the
-    /// password visible in shell history.
+    /// Use `--password-stdin` to keep the password out of shell history.
     CreateUser {
         /// Username to add or update.
         username: String,
@@ -483,8 +404,7 @@ enum GatewayCmd {
         #[arg(long)]
         password_stdin: bool,
     },
-    /// Remove a user from the gateway htpasswd file. Exits with an error
-    /// when the user is not found so callers can detect the no-op case.
+    /// Remove a user from the gateway htpasswd.
     DeleteUser {
         /// Username to remove.
         username: String,
@@ -495,19 +415,15 @@ enum GatewayCmd {
 
 #[derive(Subcommand)]
 enum WgCmd {
-    /// Generate (if absent) this hive's WireGuard private key, print its
-    /// public key, and print the nix snippet to enable the mesh. Idempotent:
-    /// an existing key is reused, never clobbered (clobbering would break a
-    /// live mesh). Share the printed public key with peer hives.
+    /// Generate this hive's WireGuard key (if absent) and print its public
+    /// key plus the nix to enable the mesh.
     Init {
         /// This hive's mesh address (e.g. `10.42.0.1/32`) to bake into the
         /// printed snippet. Omit to get a placeholder you fill in.
         #[arg(long)]
         address: Option<String>,
     },
-    /// Print the nix snippet to add a peer hive to the mesh. Pure output —
-    /// paste it into this hive's config. Get `<pubkey>` from the peer's
-    /// `hivectl wg init`.
+    /// Print the nix to add a peer hive to the mesh.
     Peer {
         /// Peer hive's DNS domain (the `swarm.peers` attrset key).
         domain: String,
@@ -522,29 +438,23 @@ enum WgCmd {
         #[arg(long)]
         endpoint: Option<String>,
     },
-    /// Show the live mesh interface state (`wg show wg-hive`). Requires the
-    /// mesh to be enabled + up.
+    /// Show the live mesh interface state.
     Status,
 }
 
 #[derive(Subcommand)]
 enum QuotaCmd {
-    /// Enable btrfs qgroup accounting on the agent-state filesystem. Run
-    /// once before `show` / `limit`. Triggers a full btrfs rescan (I/O
-    /// heavy on a large filesystem), so it's a deliberate opt-in.
-    /// Idempotent; a no-op on non-btrfs hosts.
+    /// Enable btrfs qgroup accounting on the agent-state filesystem.
+    ///
+    /// Run once before `show` / `limit`. No-op on non-btrfs hosts.
     Enable,
-    /// Report per-agent disk usage (referenced + exclusive bytes) from
-    /// btrfs qgroups. With no name, shows every agent that has a state
-    /// subvolume; pass a name to show just that one. Requires `enable`
-    /// first.
+    /// Report per-agent disk usage from btrfs qgroups (all agents, or one
+    /// by name).
     Show {
         /// Agent to show (omit for all agents with a state subvolume).
         name: Option<String>,
     },
-    /// Set or clear an agent's disk quota (a referenced-usage cap). `size`
-    /// accepts a byte count or a `K`/`M`/`G`/`T` suffix (e.g. `5G`), or
-    /// `none` to clear the limit. Requires `enable` first.
+    /// Set or clear an agent's disk-usage quota.
     Limit {
         /// Agent whose state subvolume to limit.
         name: String,
@@ -560,10 +470,7 @@ use hive_host_sock::HOST_SOCKET as DEFAULT_HOST_SOCKET;
 
 #[derive(Subcommand)]
 enum AgentsCmd {
-    /// Show all managed agents with their status (running / needs-login /
-    /// needs-update) and technical state (deployed sha, parent, pending
-    /// reminders). The host roster overview; reuses the dashboard's
-    /// per-agent aggregation. Requires the daemon running.
+    /// Show all managed agents with their status and technical state.
     List {
         /// Emit the raw JSON rows instead of the padded table (for
         /// scripting). The table is the default human-readable shape.
@@ -571,10 +478,6 @@ enum AgentsCmd {
         json: bool,
     },
     /// Stop and start a single agent container without rebuilding config.
-    /// Useful for "kick the container" when the process is stuck or the
-    /// container needs a clean restart without changing the NixOS config.
-    /// Rides the job queue (serialized against in-flight rebuilds for
-    /// the same agent); waits with live progress unless `--no-wait`.
     Restart {
         /// Agent name (e.g. `damocles`, `ruth`).
         name: String,
@@ -582,23 +485,21 @@ enum AgentsCmd {
         #[arg(long)]
         no_wait: bool,
     },
-    /// Restart ALL managed agent containers via one restart DAG each —
-    /// unrelated agents overlap, each serializes on its own lease.
-    /// Waits for the whole set with live progress unless `--no-wait`.
+    /// Restart all managed agent containers.
     RestartAll {
         /// Return immediately after the restart DAGs are queued.
         #[arg(long)]
         no_wait: bool,
     },
-    /// Spawn a new agent container directly (`h-<name>`). Bypasses the
-    /// approval queue — operator-on-the-host only. For approval-gated
-    /// spawns, use `request-spawn`.
+    /// Spawn a new agent container directly, bypassing the approval queue.
+    ///
+    /// Operator-on-the-host only; use `request-spawn` for an approval-gated
+    /// spawn.
     Spawn {
         /// Agent name (e.g. `iris`).
         name: String,
     },
-    /// Queue a spawn request as an approval. The container is created on
-    /// `hivectl approvals approve <id>` (or the dashboard APPR0VE button).
+    /// Queue a spawn request for operator approval.
     RequestSpawn {
         /// Agent name.
         name: String,
@@ -608,13 +509,12 @@ enum AgentsCmd {
         /// Agent name.
         name: String,
     },
-    /// Tear down a sub-agent container. The container is removed; persistent
-    /// state (config repos + Claude credentials) is kept by default. Pass
-    /// `--purge` to also wipe the agent's state dirs (config + creds +
-    /// notes). No undo.
+    /// Tear down a sub-agent container, keeping its state by default. No
+    /// undo.
     Destroy {
         /// Agent name.
         name: String,
+        /// Also wipe the agent's state dirs (config + creds + notes).
         #[arg(long)]
         purge: bool,
     },
@@ -623,9 +523,7 @@ enum AgentsCmd {
         /// Agent name.
         name: String,
     },
-    /// Move an agent in the topology tree. Set `--parent` to a new parent
-    /// agent name, or pass `--root` to promote the agent to root (no
-    /// parent). Exactly one is required. Refuses cycles and unknown agents.
+    /// Move an agent in the topology tree — under a new parent, or to root.
     SetParent {
         /// Agent to move.
         child: String,
@@ -638,8 +536,7 @@ enum AgentsCmd {
     },
 }
 
-/// Operator approval queue: list pending requests and approve / deny them.
-/// Requires the hive-c0re daemon (connects to the host admin socket).
+/// Operator approval queue: list, approve, or deny pending requests.
 #[derive(Subcommand)]
 enum ApprovalsCmd {
     /// List pending approval requests submitted by agents.
@@ -658,13 +555,10 @@ enum ApprovalsCmd {
 
 #[derive(Subcommand)]
 enum SubvolCmd {
-    /// Convert an existing plain-dir agent state root into a btrfs subvolume
-    /// in place. Stops the agent (so its state bind-mount is released),
-    /// migrates `…/agents/<name>/` to a subvolume preserving
-    /// ownership/permissions/xattrs, then restarts it. Idempotent (no-op if
-    /// already a subvolume) and safe (the original dir is left untouched on
-    /// any failure before the final swap). Requires `--yes` since it bounces
-    /// the agent and moves its state.
+    /// Convert a plain-dir agent state root into a btrfs subvolume in
+    /// place, so it gains snapshots and per-subvol usage/quota.
+    ///
+    /// Bounces the agent to migrate its state, so it requires `--yes`.
     Upgrade {
         /// Agent name (e.g. `damocles`, `iris`).
         name: String,
@@ -673,9 +567,7 @@ enum SubvolCmd {
         #[arg(long)]
         yes: bool,
     },
-    /// Read-only snapshots of an agent's state subvolume — the first step
-    /// of the (in-progress) inter-hive migration path, or a manual
-    /// point-in-time backup.
+    /// Read-only snapshots of an agent's state subvolume.
     Snapshot {
         #[command(subcommand)]
         cmd: SnapshotCmd,
@@ -684,8 +576,7 @@ enum SubvolCmd {
 
 #[derive(Subcommand)]
 enum SnapshotCmd {
-    /// Create a read-only snapshot. Agent must already be a subvolume
-    /// (`subvol upgrade` first). Prints the snapshot's host path.
+    /// Create a read-only snapshot (agent must already be a subvolume).
     Create {
         /// Agent name (e.g. `damocles`, `iris`).
         name: String,