fix: enable isolated live validation in no-mistakes gates

.no-mistakes.yaml

@@ -40,6 +40,8 @@ test:
     Run live Herdr scenarios only through bin/fm-herdr-lab.sh with a named non-default fm-lab-* session, following that helper's prepare, provision, run, and teardown contract exactly.
     Never touch the live default Herdr session or fleet panes.
     Prefer a throwaway lab for spawn, long-launch, and Claude-path proofs, and tear it down in the same evidence turn.
+    Use `LAB=$(bin/fm-lab-home.sh create <label>)` to create the disposable home and `bin/fm-lab-home.sh teardown "$LAB"` to remove it; the helper's header owns the marker and binding contract, and bin/fm-gate-refuse-lib.sh owns gate authorization.
+    To run a real primary inside the gate: create the lab, write the scenario's opt-in flag (e.g. `touch $LAB/config/supervision-host`), then, from the run worktree, start the harness CLI as the session command on the lab's private tmux socket: `mkdir -p $LAB/tmux && env -u NO_MISTAKES_GATE TMUX_TMPDIR=$LAB/tmux tmux -L fm-lab new-session -d -s primary -c "$PWD" -e FM_HOME=$LAB <cli>`, where <cli> is the harness's own launch command using the machine's existing login: claude -> `claude`, codex -> `codex`, cursor -> `cursor-agent`, opencode -> `opencode`, grok -> `grok`, omp -> `omp`. Drive, inspect, and stop that primary only through the same socket - `TMUX_TMPDIR=$LAB/tmux tmux -L fm-lab send-keys -t primary ...`, `TMUX_TMPDIR=$LAB/tmux tmux -L fm-lab capture-pane -p -t primary`, `TMUX_TMPDIR=$LAB/tmux tmux -L fm-lab kill-server` - never the default tmux server; the firstmate scripts the primary runs inherit $TMUX from its pane, which names that same fm-lab socket inside the lab. For a Herdr primary use a named non-default fm-lab-* session via bin/fm-herdr-lab.sh instead, and record it in the lab as soon as it is provisioned with `bin/fm-lab-home.sh record-herdr-session $LAB <session>` - the gate accepts only a Herdr session recorded for the FM_HOME's own lab. The gate marker is scrubbed (env -u NO_MISTAKES_GATE) because the lab primary is a fixture firstmate, not a gate agent. If the harness CLI is absent or its login is unavailable, report the scenario untested; never fake the CLI, the login, or the evidence.
     Do not mutate the operator primary checkout, real fleet FM_HOME state, or production credentials, and keep git changes otherwise inside the run worktree.
     Read docs/herdr-backend.md and the bin/fm-herdr-lab.sh header as the owners of Herdr lab mechanics rather than reproducing that manual here.
     Ship or scout briefs that will drive Herdr lifecycle still require --herdr-lab at scaffold time; these Test-agent instructions are not a substitute for that brief flag.

bin/fm-afk-launch.sh

@@ -190,6 +190,19 @@ fm_afk_launch_usage() {
 }
 
 fm_afk_launch_primary_harness() {
+  # Test seam: FM_TEST_HARNESS names the simulated harness for suites that
+  # invoke this launch path directly (tests/fm-afk-launch.test.sh). It is
+  # honored only alongside FM_TEST_SEAM=1, the marker test suites set, so a
+  # leaked variable in a real primary's environment stays inert. Only an
+  # exact known-harness token is honored, so a stray or mistyped value falls
+  # through to real detection; bin/fm-harness.sh's production detect_own
+  # precedence never reads either variable.
+  case "${FM_TEST_SEAM:-}${FM_TEST_SEAM:+ }${FM_TEST_HARNESS:-}" in
+    "1 claude" | "1 codex" | "1 opencode" | "1 pi" | "1 pi-signed" | "1 grok" | "1 kimi" | "1 cursor" | "1 gemini" | "1 muse" | "1 rovo" | "1 omp" | "1 agy" | "1 devin" | "1 unknown")
+      printf '%s' "$FM_TEST_HARNESS"
+      return
+      ;;
+  esac
   "$FM_AFK_LAUNCH_DIR/fm-harness.sh" 2>/dev/null || printf unknown
 }
 

bin/fm-control.sh

@@ -145,9 +145,10 @@ esac
 
 # shellcheck source=bin/fm-gate-refuse-lib.sh
 . "$SCRIPT_DIR/fm-gate-refuse-lib.sh"
-# Fail closed before any fleet mutation: a no-mistakes gate agent must never
-# drive a crewmate's lifecycle (see bin/fm-gate-refuse-lib.sh).
-fm_refuse_if_gate_agent
+# Refuse gate-context lifecycle calls unless the target is a verified lab
+# (see bin/fm-gate-refuse-lib.sh). The backend comes from recorded task
+# metadata, not the environment, so the env-backend check is skipped.
+fm_refuse_if_gate_agent . '' 1
 
 if [ -z "${FM_HOME+x}" ] || [ -z "${FM_HOME:-}" ]; then
   echo "error: FM_HOME is not set; fm-control refuses to resolve a task without an explicit firstmate home" >&2
@@ -341,6 +342,8 @@ fi
 fm_backend_validate_task_endpoint "$META" "$ID" || exit 1
 BACKEND=$FM_BACKEND_VALIDATED_BACKEND
 T=$FM_BACKEND_VALIDATED_TARGET
+fm_gate_lab_assert_backend "$BACKEND"
+fm_gate_lab_assert_target "$BACKEND" "$T"
 LABEL="fm-$ID"
 RECORDED_HARNESS=$(fm_meta_get "$META" harness)
 KIND=$(fm_meta_get "$META" kind)