test: cover machine template registry

README.md

@@ -401,6 +401,34 @@ unattended VM-host appliance) — and `workstation` is the machine at the keyboa
 end of all the SSH connections: `root-door=closed`, `join=login`, entering the
 tailnet as *your* device rather than the fleet's.
 
+### Machine-role templates
+
+Machine presets can also live in the
+[heavy-duty/rig-templates](https://github.com/heavy-duty/rig-templates)
+registry. A `*-server` directory is a fleet-machine definition; the exact
+name `workstation` is the deliberate suffix-less exception. Its
+`template.env` contains exactly the three traits bootstrap's built-in table
+uses:
+
+```dotenv
+ROOT_DOOR="open"    # open|closed
+HOST="no"           # yes|no
+JOIN="authkey"      # authkey|login
+```
+
+An `install.sh` is optional. When present, bootstrap runs it as root,
+non-interactively, from the definition directory with `RIG_ROLE` set, after
+the tailnet join, host setup, role marker prerequisites, and operator
+convergence. A nonzero exit fails bootstrap and names the role and registry
+source. The definition owns idempotence, just as bootstrap does.
+
+Built-in roles and `custom` take precedence over registry names. Any other
+non-tenant role is looked up in the resolved registry; the same three source
+knobs below apply, including `RIG_TEMPLATES_DIR` for an offline local
+definition. Pin reviewed registry content into rig's tree before using it on
+fleet machines: an optional machine `install.sh` executes as root on metal,
+and an override is the operator explicitly choosing a different trust root.
+
 ### `rig bootstrap <role>-box` — the box tenants
 
 Run as root, **inside** a [box](https://github.com/heavy-duty/box)-minted

changelog.d/152.md

@@ -0,0 +1 @@
+Machine-role templates can declare bootstrap traits and an optional final root install hook (#152)

commands/bootstrap.sh

@@ -144,7 +144,8 @@ case "$ROLE" in
     templates_resolve || exit 2
     trap '[ -n "$TEMPLATES_TMP" ] && rm -rf "$TEMPLATES_TMP"' EXIT
     MACHINE_TEMPLATE_DIR="$REGISTRY_DIR/$ROLE"
-    if [ "$(template_family "$ROLE" 2>/dev/null || true)" != "machine" ] \
+    if { [ "$ROLE" != "workstation" ] && [[ ! "$ROLE" =~ ^[a-z][a-z0-9-]*-server$ ]]; } \
+      || [ "$(template_family "$ROLE" 2>/dev/null || true)" != "machine" ] \
       || [ ! -f "$MACHINE_TEMPLATE_DIR/template.env" ]; then
       MACHINE_ROLES="$(templates_machine_roles "$REGISTRY_DIR" | paste -sd'|' -)"
       [ -n "$MACHINE_ROLES" ] || MACHINE_ROLES="none"

commands/template-lint.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env bash
-# rig template-lint <role-dir>... — is this a valid tenant-role definition?
+# rig template-lint <role-dir>... — is this a valid role definition?
 #
 # rig defines what a valid template is (the schema lives in
 # lib/templates.sh, beside the mint-time parser that enforces it); the
@@ -23,12 +23,15 @@ usage() {
   cat <<'EOF'
 usage: rig template-lint <role-dir>...
 
-Validate tenant-role definitions (the heavy-duty/rig-templates shape):
+Validate role definitions (the heavy-duty/rig-templates shape).
-each <role-dir> must carry a family-suffixed name (rig#76), a template.env
+
-that parses against the allowlist (KEY="value" only — the file is data,
+Tenant roles use a *-box directory, tenant template.env schema, a shebang
-never sourced), an install.sh with a shebang, and a non-blank creds.md.
+install.sh, and non-blank creds.md. Machine roles use a *-server directory
-Every refusal names the failing key or file. Exits non-zero if any
+(or exact name workstation), the ROOT_DOOR/HOST/JOIN schema, no creds.md,
-definition fails; nothing is written.
+and an optional install.sh which must be non-empty and carry a shebang.
+template.env is parsed as KEY="value" data and never sourced. Every refusal
+names the failing key or file. Exits non-zero if any definition fails;
+nothing is written.
 EOF
 }
 

test/cli.sh

@@ -613,6 +613,25 @@ mkdir -p "$TPL_FIX/badnode-box"
 printf 'USER="x"\nCONTEXT_PATH=".x/A.md"\nCLI_NAME="x"\nPATH_LINE="p"\nNEEDS_NODE="maybe"\n' > "$TPL_FIX/badnode-box/template.env"
 mkdir -p "$TPL_FIX/badapt-box"
 printf 'USER="x"\nCONTEXT_PATH=".x/A.md"\nCLI_NAME="x"\nPATH_LINE="p"\nAPT_EXTRAS="zsh -o"\n' > "$TPL_FIX/badapt-box/template.env"
+mkdir -p "$TPL_FIX/scratch-server"
+printf 'ROOT_DOOR="closed"\nHOST="no"\nJOIN="login"\n' > "$TPL_FIX/scratch-server/template.env"
+mkdir -p "$TPL_FIX/workstation"
+printf 'ROOT_DOOR="closed"\nHOST="yes"\nJOIN="login"\n' > "$TPL_FIX/workstation/template.env"
+mkdir -p "$TPL_FIX/hooked-server"
+printf 'ROOT_DOOR="open"\nHOST="no"\nJOIN="authkey"\n' > "$TPL_FIX/hooked-server/template.env"
+printf '#!/usr/bin/env bash\nexit 1\n' > "$TPL_FIX/hooked-server/install.sh"
+mkdir -p "$TPL_FIX/baddoor-server"
+printf 'ROOT_DOOR="ajar"\nHOST="no"\nJOIN="authkey"\n' > "$TPL_FIX/baddoor-server/template.env"
+mkdir -p "$TPL_FIX/tenantkeys-server"
+cp "$TPL_FIX/scratch-box/template.env" "$TPL_FIX/tenantkeys-server/template.env"
+mkdir -p "$TPL_FIX/machinekeys-box"
+cp "$TPL_FIX/scratch-server/template.env" "$TPL_FIX/machinekeys-box/template.env"
+mkdir -p "$TPL_FIX/creds-server"
+cp "$TPL_FIX/scratch-server/template.env" "$TPL_FIX/creds-server/template.env"
+printf 'not used\n' > "$TPL_FIX/creds-server/creds.md"
+mkdir -p "$TPL_FIX/noshebang-server"
+cp "$TPL_FIX/scratch-server/template.env" "$TPL_FIX/noshebang-server/template.env"
+printf 'exit 0\n' > "$TPL_FIX/noshebang-server/install.sh"
 
 # THE HARD CUT, tenant half (#76). The pre-rename names are gone and must fail
 # as UNKNOWN — asserted per name, because an alias left in for one tenant is the
@@ -653,6 +672,28 @@ check "bootstrap: tenant roles dispatch through bootstrap.sh" 0 "Box TENANT role
 check "bootstrap: an unheard-of '-box' role still dispatches (zero code changes)" 0 "Box TENANT roles" \
   "$ROOT/commands/bootstrap.sh" scratch-box --help
 
+# Machine roles use the same resolved registry but remain table-compatible:
+# loading happens before flag parsing, so an explicit flag overrides the
+# definition exactly as it overrides a built-in row.
+check "machine template: traits load from the local registry" 2 "join=login" \
+  env RIG_TEMPLATES_DIR="$TPL_FIX" TS_AUTHKEY=x \
+  "$ROOT/commands/bootstrap.sh" scratch-server --no-users
+if [ "$(id -u)" -ne 0 ]; then
+  check "machine template: a flag overrides the loaded trait" 1 "must run as root" \
+    env RIG_TEMPLATES_DIR="$TPL_FIX" TS_AUTHKEY=x \
+    "$ROOT/commands/bootstrap.sh" scratch-server --no-users --join authkey
+fi
+check "machine template: invalid ROOT_DOOR is refused by key" 2 "ROOT_DOOR" \
+  env RIG_TEMPLATES_DIR="$TPL_FIX" \
+  "$ROOT/commands/bootstrap.sh" baddoor-server --no-users
+check "machine template: unknown role lists machine definitions" 2 "scratch-server" \
+  env RIG_TEMPLATES_DIR="$TPL_FIX" "$ROOT/commands/bootstrap.sh" absent-server
+check "machine template: unknown role names the resolved source" 2 "RIG_TEMPLATES_DIR" \
+  env RIG_TEMPLATES_DIR="$TPL_FIX" "$ROOT/commands/bootstrap.sh" absent-server
+check "machine template: a registry role cannot shadow a built-in" 2 "unset TS_AUTHKEY" \
+  env RIG_TEMPLATES_DIR="$TPL_FIX" TS_AUTHKEY=x \
+  "$ROOT/commands/bootstrap.sh" workstation --no-users
+
 # The tenant marker guard (#83), against marker FIXTURES (never the harness
 # machine's real /etc/rig/role): converging a tenant onto a machine-role box or a
 # VM host (host=yes) refuses for every tenant — and names the staging PAIR,
@@ -844,6 +885,30 @@ cp "$TPL_FIX/scratch-box/template.env" "$TPL_FIX/scratch-box/creds.md" "$TPL_FIX
 printf 'exit 0\n' > "$TPL_FIX/noshebang-box/install.sh"
 check "template-lint: an install.sh without a shebang is refused" 1 "no shebang" \
   "$ROOT/commands/template-lint.sh" "$TPL_FIX/noshebang-box"
+check "template-lint: a traits-only machine definition passes" 0 "OK: " \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/scratch-server"
+check "template-lint: workstation is the machine-family carve-out" 0 "OK: " \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/workstation"
+check "template-lint: machine roles refuse tenant keys" 1 "unknown key: USER" \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/tenantkeys-server"
+check "template-lint: tenant roles refuse machine keys" 1 "unknown key: ROOT_DOOR" \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/machinekeys-box"
+check "template-lint: machine roles refuse creds.md" 1 "creds.md is not allowed" \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/creds-server"
+check "template-lint: machine install.sh requires a shebang" 1 "no shebang" \
+  "$ROOT/commands/template-lint.sh" "$TPL_FIX/noshebang-server"
+# The install is deliberately after the users phase and its wrapper names both
+# role and source. Dynamic execution belongs to the root integration path; the
+# non-root offline harness pins the safety ordering and failure contract.
+machine_hook_at="$(grep -n 'running install hook for' "$ROOT/commands/bootstrap.sh" | head -n1 | cut -d: -f1)"
+check "machine template: install hook is bootstrap's last convergence phase" 0 "" \
+  test "${users_apply_at:-999999}" -lt "${machine_hook_at:-0}"
+# shellcheck disable=SC2016
+check "machine template: install failure names role and source" 0 "" \
+  grep -qF 'install hook failed for role $ROLE from $(templates_source_desc)' "$ROOT/commands/bootstrap.sh"
+# shellcheck disable=SC2016
+check "machine template: install runs from its definition with RIG_ROLE" 0 "" \
+  grep -qF 'cd "$MACHINE_TEMPLATE_DIR" && RIG_ROLE="$ROLE" ./install.sh' "$ROOT/commands/bootstrap.sh"
 rm -rf "$TPL_FIX" "$TPL_WORK"
 
 # Creds-free BY CONSTRUCTION, provable by absence (box#69's grep-refusal