docs: the handbook follows the three layers: swarm, hives, agents

The handbook's sections were agent-lifecycle, integrations, swarm, tools,
trust-boundary and web-ui, which do not follow how the system is built, so a
reader could not tell which layer a page is about. Its pages now sit in
swarm/, hives/ and agents/, with the accounts an agent can hold outside the
system in external-accounts/.

- swarm/: forge, matrix and the knowledge repo join the swarm pages, and so
  does the swarmctl reference.
- hives/: the dashboard, hivectl and its reference, and the operator/agent
  boundary, with a README that takes over the old web-ui and tools pages.
- agents/: the roster, persistence, config changes, the agent's own web page
  (was web-ui/agent) and the security model, with a README.
- external-accounts/: GitHub and GitLab.

Pages are moved whole; none is split yet. Every reference follows: relative
links in the docs, `docs/2-handbook/...` paths in code comments, the nix
checks that regenerate the CLI references and the ignore list, and docs/README.md,
whose handbook section is now by layer.

hive-forge-notify/src/notify.rs

@@ -18,12 +18,12 @@
 //! **Multi-source**: always the internal Forgejo, plus github.com when the
 //! agent has a PAT. Each source polls independently behind
 //! [`Source`]; everything below is shared. Rationale
-//! + host differences: `docs/2-handbook/integrations/github.md::Notifications`.
+//! + host differences: `docs/2-handbook/external-accounts/github.md::Notifications`.
 //!
 //! Activation gates, self-notification filtering, body excerpt +
 //! truncation + heading escape, wrapper formats (comment / review /
 //! new-item / state-change), meta suffix, and review-request override
-//! all live in `docs/2-handbook/integrations/forge.md::Notification poller`.
+//! all live in `docs/2-handbook/swarm/forge.md::Notification poller`.
 
 use std::collections::{HashMap, HashSet};
 use std::fmt::Write as _;
@@ -85,7 +85,7 @@ where
 /// a `new <kind>` label. Later activity that lands on the state-change
 /// path because it carries no `latest_comment_url` (e.g. a bodiless
 /// review submission) fires well outside this window, so we must not
-/// claim it's "new" — see docs/2-handbook/integrations/forge.md, "new vs activity on".
+/// claim it's "new" — see docs/2-handbook/swarm/forge.md, "new vs activity on".
 const NEW_ITEM_TOLERANCE_SECS: i64 = 120;
 
 /// Fetch the account's own login for self-notification filtering.
@@ -133,7 +133,7 @@ fn notif_type_label(t: Option<NotifySubjectType>) -> &'static str {
 /// Escape strict-ATX markdown headings in a body before embedding it
 /// inside the forge-notify wrapper, so a leading `## title` line
 /// doesn't blow into an h2 in the dashboard render. See
-/// `docs/2-handbook/integrations/forge.md::Body excerpt + truncation + heading escape` for
+/// `docs/2-handbook/swarm/forge.md::Body excerpt + truncation + heading escape` for
 /// the strict-ATX-vs-`#tag` rationale and the `split_inclusive`
 /// trailing-newline contract.
 fn escape_md_headings(body: &str) -> String {
@@ -218,7 +218,7 @@ fn is_username_byte(b: u8) -> bool {
 /// `@username` mention AND aren't already present (as a substring) in
 /// `included_excerpt`. Surfaces tags that fell outside the truncation
 /// window so addressed agents never silently miss a mention on a long
-/// body. See `docs/2-handbook/integrations/forge.md::Body excerpt + truncation + heading
+/// body. See `docs/2-handbook/swarm/forge.md::Body excerpt + truncation + heading
 /// escape` for the truncate-before-escape ordering rule.
 fn extract_truncated_mention_lines<'a>(full_body: &'a str, included_excerpt: &str) -> Vec<&'a str> {
     full_body
@@ -275,7 +275,7 @@ fn review_state_label(state: &str) -> Option<&str> {
 /// Build a human-readable wake message for one Forgejo notification,
 /// or `None` for a self-echo the caller should mark-read without
 /// delivery. Wrapper format table + meta-suffix shape + number/repo
-/// extraction live in `docs/2-handbook/integrations/forge.md::Wrapper format` +
+/// extraction live in `docs/2-handbook/swarm/forge.md::Wrapper format` +
 /// `::Meta suffix`.
 async fn format_notification<S: Source>(
     client: &reqwest::Client,
@@ -327,7 +327,7 @@ async fn format_notification<S: Source>(
 
     // Always fetch subject detail for assignee/reviewer metadata so
     // the meta suffix can show current ownership without a follow-up
-    // fetch (see `docs/2-handbook/integrations/forge.md::Meta suffix`).
+    // fetch (see `docs/2-handbook/swarm/forge.md::Meta suffix`).
     let subject = if subject_api_url.is_empty() {
         None
     } else {
@@ -417,7 +417,7 @@ async fn format_notification<S: Source>(
 /// that comment was genuinely posted AFTER the close - a comment
 /// racing the merge inside the state-change tolerance window - it is
 /// appended so it isn't lost (best of both worlds; see
-/// `docs/2-handbook/integrations/forge.md::Merge racing a comment`). The ordinary
+/// `docs/2-handbook/swarm/forge.md::Merge racing a comment`). The ordinary
 /// pre-merge last comment (created before `closed_at`) is left off.
 /// The earlier-updates count is the same widened marker as the comment
 /// path; both are fetched here because `format_state_change_notification`
@@ -463,7 +463,7 @@ struct NotifMeta<'a> {
 
 /// Build the `\nassignee: ...` (and optionally `\nreviewer: ...`)
 /// suffix appended to every wrapper. Shape + presence rules live in
-/// `docs/2-handbook/integrations/forge.md::Meta suffix`.
+/// `docs/2-handbook/swarm/forge.md::Meta suffix`.
 fn build_meta_suffix(subject: Option<&serde_json::Value>, is_pr: bool) -> String {
     let assignees: Vec<&str> = subject
         .and_then(|s| s["assignees"].as_array())
@@ -725,7 +725,7 @@ fn format_state_change_notification(
     // time. A review submitted with no body carries no
     // `latest_comment_url`, so it lands here instead of on the comment
     // path — and its event time is well after `created_at`. Labeling
-    // that "new PR" is misleading (see docs/2-handbook/integrations/forge.md, "new vs activity
+    // that "new PR" is misleading (see docs/2-handbook/swarm/forge.md, "new vs activity
     // on"): agents dismiss it as a
     // duplicate of the original open notification. When we can't confirm
     // creation, fall back to a neutral "activity on" label.
@@ -757,7 +757,7 @@ fn format_state_change_notification(
     // Review-request override: Forgejo doesn't reliably set
     // `reason == "review_requested"` (often null), so we check the
     // subject payload's `requested_reviewers` list directly. See
-    // `docs/2-handbook/integrations/forge.md::Review-request override`.
+    // `docs/2-handbook/swarm/forge.md::Review-request override`.
     let is_review_request = is_new
         && *is_pr
         && !own_login.is_empty()
@@ -801,7 +801,7 @@ fn format_state_change_notification(
 /// `created_at`: within `NEW_ITEM_TOLERANCE_SECS` ⇒ creation. When
 /// either timestamp is missing or unparseable we default to `true`,
 /// preserving the prior "new" behavior rather than masking a genuine
-/// new item behind the neutral fallback. See docs/2-handbook/integrations/forge.md, "new vs
+/// new item behind the neutral fallback. See docs/2-handbook/swarm/forge.md, "new vs
 /// activity on".
 fn notification_is_creation(
     event: Option<OffsetDateTime>,