docs(claude): mention top-level scripts/ in Layout (#3)

Add a § Layout entry for the new top-level scripts/ directory and
its four subdirectories (driver, watcher, daemons, systemd). Per
designs/driver.md § Modified artifacts, no `driver` row is added to
the Current inventory's role list because the driver is a script,
not a role; the new skill files (driver-pr-creation-state-machine,
prompt-on-failure-capture, activity-feed-watcher) are appended to
the skills inventory alongside the previously-unlisted ones already
landed by sibling PRs (cleaner, gardener-inbox-error-reporting,
driver-design-only-pr-workflow).

Add .gitignore entry for scripts/daemons/config.sh (host-local
maintainer-edited driver-pool and watcher-feed configuration; the
.example template is checked in, the actual config is not).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: endolinbot

.gitignore

@@ -18,3 +18,8 @@
 !/.gitignore
 !/.github/
 !/.github/**
+
+# Host-local daemon configuration for the driver pool and per-feed
+# watchers. The .example template is checked in; the actual config.sh
+# is host-local. See scripts/daemons/README.md.
+/scripts/daemons/config.sh

CLAUDE.md

@@ -1,7 +1,7 @@
 ---
 created: 2026-05-12
-updated: 2026-05-15
-author: gardener, liaison
+updated: 2026-06-02
+author: gardener, liaison, builder
 ---
 
 
@@ -20,6 +20,7 @@ The garden is a library of agent **roles** and **skills** for working across man
 - `journal/`: a git worktree of this repo on the orphan branch `journal`. Holds the garden's transcript and acts as the **two-channel message bus** between agents: a per-role inbox (`journal/inboxes/<host>/<role>.md`; drained via `skills/inbox-drain/SKILL.md`) for directed communication, and a **job board** (`journal/jobs/`; contract at [`journal/jobs/README.md`](journal/jobs/README.md); skill at [`skills/job-board/SKILL.md`](skills/job-board/SKILL.md)) for work items that any eligible consumer can race to claim via git push as the serialization point. See [WORKTREES.md](WORKTREES.md) for the worktree shapes.
 - `worktrees/<owner>-<repo>.git/`: bare clones of upstream forks.
 - `worktrees/<owner>-<repo>/<name>/`: fork worktrees the garden is currently working in. Naming and lifecycle in [WORKTREES.md](WORKTREES.md).
+- `scripts/`: executable shell scripts for humans and systemd. Holds the driver (`scripts/driver/`), the per-feed activity watchers (`scripts/watcher/<feed>/`), the daemon-management wrappers (`scripts/daemons/`), and the templated systemd user units (`scripts/systemd/`). The split is strict: `roles/` and `skills/` hold no executables; `scripts/` holds no agent-only context fragments. See [designs/driver.md](designs/driver.md) for the rationale.
 - `references/`: read-only shelves of roles and skills imported from other gardens. Browsed by the liaison when a user prompt has no obvious fit in the active library, never auto-loaded by subagents. See [references/README.md](references/README.md).
 
 Files are named `AGENT.md` / `SKILL.md` / `COMMON.md` (not `CLAUDE.md`) on purpose: we do **not** want Claude Code to auto-load them into a subagent's context. They are loaded explicitly by the dispatched subagent.
@@ -140,6 +141,6 @@ The constraint covers both **event-level** surveillance (the standing-monitor da
 ## Current inventory
 
 - Roles: `liaison`, `steward`, `general-contractor`, `monitor`, `review-queue`, `boatman`, `builder`, `assayer`, `cleaner`, `solicitor`, `barrister`, `justice`, `appellate`, `fixer`, `weaver`, `shepherd`, `conductor`, `designer`, `scout`, `botanist`, `major-general`, `gardener`, `evaluator`, `groom`, `investigator`, `journalist`, `librarian`, `scholar`, `timekeeper`. The prior single `judge` role split into three on 2026-05-21: `solicitor` (designer work; design panel), `barrister` (builder work; first code-panel round), `justice` (fixer work; code-panel re-runs); `roles/judge/AGENT.md` is retained as a redirect. The `appellate` lands the same day to appeal `follow-up` and `acknowledge` dispositions on small-and-in-context items into `summary-fix` before un-draft. Plus the jury-seat roles dispatched by the three judges: a code panel for source-touching PRs and a design panel for design-only PRs (paths under `<project>/designs/`). See each judge's role file under `roles/<solicitor|barrister|justice>/AGENT.md` for the panel composition; the orchestrator never dispatches a juror seat directly.
-- Skills: `journal-sync`, `self-improvement`, `em-dash-style`, `relative-paths`, `agent-termination`, `rule-elision-test`, `inbox-drain`, `autonomous-loop-pacing`, `github-activity-poll`, `pr-ci-watch`, `review-queue-poll`, `rebase-before-followup`, `review-feedback-followup-commits`, `pr-review-thread-replies`, `pr-formation`, `pr-handoff`, `pr-creation-flow`, `pr-dependency-graph`, `pr-dependency-topo-sort`, `yarn-lock-separate-commit`, `pre-pr-checklist`, `regression-evidence`, `coverage-driven-testing`, `adversarial-tests`, `saboteur-adversarial-review`, `panel-review`, `ci-status-summary`, `ci-runtime-comparison`, `conflict-resolution`, `cherry-pick-followup`, `rebase-hygiene-audit`, `worktree-per-pr`, `process-documents`, `prompt-section-discovery`, `benchmark-comparative-report`, `verify-upstream-state-before-pinning`, `reactji-acknowledgment`, `changeset-discipline`, `rename-discipline`, `monitor-arming`, `context-library`, `journalism`, `dispatch-worktree`, `scheduling`, `velocity-recalibration`, `roadmap-projection`, `dependency-graph-maintenance`, `groom-open-questions`, `design-queue-drift-check`, `design-to-pr-pipeline`, `garden-ab-evaluation`, `merged-pr-feedback-watch`, `library-lookup`, `retcon`, `design-dependency-walk`, `stacked-pr-build`, `gap-revealing-build`, `at-mention-surveillance`, `job-board`, `pre-push-gates`, `node-lts-window-watch`, `panel-hints`, `frozen-base-branch`. Per-project monitor reaction skills (`monitor-endo`, `monitor-endo-but-for-bots`, `monitor-agoric-sdk`, `monitor-cosgov`, `monitor-garden`) live alongside but are configuration for the `monitor` role rather than independently reusable procedures. `monitor-garden` is the only one whose dispatched subagent runs as `liaison` rather than `monitor`; see that skill's *Dispatch role asymmetry* for why.
+- Skills: `journal-sync`, `self-improvement`, `em-dash-style`, `relative-paths`, `agent-termination`, `rule-elision-test`, `inbox-drain`, `autonomous-loop-pacing`, `github-activity-poll`, `pr-ci-watch`, `review-queue-poll`, `rebase-before-followup`, `review-feedback-followup-commits`, `pr-review-thread-replies`, `pr-formation`, `pr-handoff`, `pr-creation-flow`, `pr-dependency-graph`, `pr-dependency-topo-sort`, `yarn-lock-separate-commit`, `pre-pr-checklist`, `regression-evidence`, `coverage-driven-testing`, `adversarial-tests`, `saboteur-adversarial-review`, `panel-review`, `ci-status-summary`, `ci-runtime-comparison`, `conflict-resolution`, `cherry-pick-followup`, `rebase-hygiene-audit`, `worktree-per-pr`, `process-documents`, `prompt-section-discovery`, `benchmark-comparative-report`, `verify-upstream-state-before-pinning`, `reactji-acknowledgment`, `changeset-discipline`, `rename-discipline`, `monitor-arming`, `context-library`, `journalism`, `dispatch-worktree`, `scheduling`, `velocity-recalibration`, `roadmap-projection`, `dependency-graph-maintenance`, `groom-open-questions`, `design-queue-drift-check`, `design-to-pr-pipeline`, `garden-ab-evaluation`, `merged-pr-feedback-watch`, `library-lookup`, `retcon`, `design-dependency-walk`, `stacked-pr-build`, `gap-revealing-build`, `at-mention-surveillance`, `job-board`, `pre-push-gates`, `node-lts-window-watch`, `panel-hints`, `frozen-base-branch`, `cleaner`, `gardener-inbox-error-reporting`, `driver-pr-creation-state-machine`, `driver-design-only-pr-workflow`, `prompt-on-failure-capture`, `activity-feed-watcher`. The last six are the script-orchestrated PR-creation flow's agent context (per [designs/driver.md](designs/driver.md)); the executable counterparts live under `scripts/`. Per-project monitor reaction skills (`monitor-endo`, `monitor-endo-but-for-bots`, `monitor-agoric-sdk`, `monitor-cosgov`, `monitor-garden`) live alongside but are configuration for the `monitor` role rather than independently reusable procedures. `monitor-garden` is the only one whose dispatched subagent runs as `liaison` rather than `monitor`; see that skill's *Dispatch role asymmetry* for why.
 
 The `liaison` and `steward` are the two top-level orchestrator postures. When a user is in the loop (this terminal session), the liaison runs with excess authority and asks before acting. When the garden runs in the bot sandbox under safe bot credentials with no user present, the steward runs with bounded authority and may act on its own. Multiple stewards can run concurrently across hosts (and even within one host) and share work load via the journal's job board (`journal/jobs/`); the claim race resolves contention without a dedicated peer-role posture. A third posture, the [`general-contractor`](roles/general-contractor/AGENT.md), is a liaison-adopted, focused, parallelized PR-pipeline orchestrator that maintains up to three concurrent PR-pipeline slots; the maintainer names a liaison session as the contractor when the focused, parallel design-to-PR work is wanted in the foreground. The contractor holds the steward's bounded authority bounds. See `roles/liaison/AGENT.md` § Posture, `roles/steward/AGENT.md` § Posture, and `roles/general-contractor/AGENT.md` § Posture for the three-posture contract.