feat(run): critic owns the deadline — MaxRuntime becomes the soft trigger

When a run enables the critic (Ports.Critic set + RunnableAgent.Critic.Enabled),
the kernel no longer hard-caps it at MaxRuntime. MaxRuntime becomes the SOFT
trigger (passed to startCritic, used by the host critic as its wake + the base
for its extendable backstop); the critic's deadline-watch is the real hard
cancel. This restores mort's old agentexec two-tier timeout semantics — a
slow-but-progressing run (e.g. a parent agent blocked on a 30-min animate render)
is given room up to the critic's backstop instead of being killed at the nominal
MaxRuntime.

Specifics:
- run/executor.go: the WithTimeout(MaxRuntime) is now conditional. Non-critic
  runs keep the literal MaxRuntime kill (→ "timeout"). Critic-owned runs get a
  GENEROUS WithTimeout at the new Defaults.CriticAbsoluteMax (default 6h) as a
  failsafe ceiling only — it never fires before the critic's backstop, and it
  guarantees a broken/nil host handle can't run unbounded.
- run/critic.go: startCritic takes the resolved MaxRuntime as the soft trigger
  (falling back to Defaults.CriticSoftTimeout, then 90s), instead of always using
  the global CriticSoftTimeout.
- Defaults.CriticAbsoluteMax added (withFallbacks default 6h).
- Tests: non-critic dies at MaxRuntime; critic-owned survives past it; soft
  trigger == MaxRuntime.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Jo75sqmeVPgFUWZQBn179X

run/critic.go

@@ -31,12 +31,20 @@ type criticBinding struct {
 // "killed"); when the backstop simply expired, it is context.DeadlineExceeded (→
 // "timeout"). Returns (nil, no-op stop) when there is no critic. The caller MUST
 // defer the returned stop.
-func (e *Executor) startCritic(runCtx context.Context, cancelCause context.CancelCauseFunc, ra RunnableAgent, info RunInfo) (*criticBinding, func()) {
+//
+// softTrigger is the run's resolved MaxRuntime: for a critic-owned run MaxRuntime
+// is the soft wake (mort's two-tier semantics — the critic first reviews once the
+// run exceeds its nominal budget, and its backstop = softTrigger × multiplier).
+// It falls back to the configured CriticSoftTimeout when the run set no MaxRuntime.
+func (e *Executor) startCritic(runCtx context.Context, cancelCause context.CancelCauseFunc, ra RunnableAgent, info RunInfo, softTrigger time.Duration) (*criticBinding, func()) {
 	noop := func() {}
 	if e.cfg.Ports.Critic == nil || !ra.Critic.Enabled {
 		return nil, noop
 	}
-	soft := e.cfg.Defaults.CriticSoftTimeout
+	soft := softTrigger
+	if soft <= 0 {
+		soft = e.cfg.Defaults.CriticSoftTimeout
+	}
 	if soft <= 0 {
 		soft = 90 * time.Second // defensive: withFallbacks normally guarantees >0
 	}

run/critic_deadline_test.go

@@ -0,0 +1,131 @@
+package run_test
+
+import (
+	"context"
+	"sync"
+	"testing"
+	"time"
+
+	"gitea.stevedudenhoeffer.com/steve/majordomo/llm"
+	"gitea.stevedudenhoeffer.com/steve/majordomo/provider/fake"
+
+	"gitea.stevedudenhoeffer.com/steve/executus/run"
+	"gitea.stevedudenhoeffer.com/steve/executus/tool"
+)
+
+// slowToolInvocation builds an Invocation whose session factory adds a "slow"
+// tool that sleeps for d (respecting ctx). The model script calls it once, then
+// answers — so the run's wall-clock is dominated by d, letting a test set a tiny
+// MaxRuntime and observe whether MaxRuntime hard-cancels the run.
+func slowToolInvocation(runID string, d time.Duration) tool.Invocation {
+	slow := llm.DefineTool("slow", "sleeps for a while",
+		func(ctx context.Context, _ struct{}) (any, error) {
+			select {
+			case <-time.After(d):
+				return "ok", nil
+			case <-ctx.Done():
+				return nil, ctx.Err()
+			}
+		})
+	return tool.Invocation{
+		RunID: runID,
+		SessionToolFactory: func(_ tool.AgentSession) tool.SessionTools {
+			return tool.SessionTools{Tools: []llm.Tool{slow}}
+		},
+	}
+}
+
+func slowModel() llm.Model {
+	fp := fake.New("fake")
+	fp.Enqueue("m",
+		fake.ReplyWith(llm.Response{ToolCalls: []llm.ToolCall{{ID: "c1", Name: "slow", Arguments: []byte(`{}`)}}}),
+		fake.Reply("done"),
+	)
+	m, _ := fp.Model("m")
+	return m
+}
+
+// TestNoCritic_MaxRuntimeIsHardCap: the legacy contract is preserved — without a
+// critic, MaxRuntime is a literal WithTimeout that kills a run whose work outlasts
+// it. The slow tool (200ms) outlasts MaxRuntime (20ms), so runCtx cancels mid-tool
+// and the run ends in error (timeout).
+func TestNoCritic_MaxRuntimeIsHardCap(t *testing.T) {
+	m := slowModel()
+	ex := run.New(run.Config{
+		Registry: tool.NewRegistry(),
+		Models:   func(ctx context.Context, _ string) (context.Context, llm.Model, error) { return ctx, m, nil },
+	})
+	res := ex.Run(context.Background(),
+		run.RunnableAgent{Name: "x", ModelTier: "m", MaxIterations: 5, MaxRuntime: 20 * time.Millisecond},
+		slowToolInvocation("r", 200*time.Millisecond), "go")
+	if res.Err == nil {
+		t.Fatalf("non-critic run should hard-timeout at MaxRuntime; got output=%q err=nil", res.Output)
+	}
+}
+
+// TestCriticOwnsDeadline_SurvivesPastMaxRuntime: the fix — when the critic owns the
+// deadline (Ports.Critic set + Critic.Enabled), MaxRuntime becomes the SOFT trigger
+// and is NOT a hard cap. The fake critic exposes no hard deadline (Deadline()==zero,
+// no kill), so the only hard ceiling is CriticAbsoluteMax (10s here). The slow tool
+// (200ms) outlasts the tiny MaxRuntime (20ms) but the run completes — proving the
+// old agentexec two-tier semantics are restored.
+func TestCriticOwnsDeadline_SurvivesPastMaxRuntime(t *testing.T) {
+	m := slowModel()
+	h := &fakeCriticHandle{} // Deadline()==zero → no hard deadline, no kill
+	ex := run.New(run.Config{
+		Registry: tool.NewRegistry(),
+		Models:   func(ctx context.Context, _ string) (context.Context, llm.Model, error) { return ctx, m, nil },
+		Ports:    run.Ports{Critic: &fakeCritic{h: h}},
+		Defaults: run.Defaults{CriticAbsoluteMax: 10 * time.Second},
+	})
+	res := ex.Run(context.Background(),
+		run.RunnableAgent{Name: "watched", ModelTier: "m", MaxIterations: 5, MaxRuntime: 20 * time.Millisecond,
+			Critic: run.CriticConfig{Enabled: true}},
+		slowToolInvocation("r", 200*time.Millisecond), "go")
+	if res.Err != nil {
+		t.Fatalf("critic-owned run must survive past MaxRuntime (soft trigger); got err=%v", res.Err)
+	}
+	if res.Output != "done" {
+		t.Errorf("output = %q, want %q", res.Output, "done")
+	}
+}
+
+// capturingCritic records the soft trigger the executor passes to Monitor.
+type capturingCritic struct {
+	mu   sync.Mutex
+	soft time.Duration
+	h    run.CriticHandle
+}
+
+func (c *capturingCritic) Monitor(_ context.Context, _ run.RunInfo, soft time.Duration) run.CriticHandle {
+	c.mu.Lock()
+	c.soft = soft
+	c.mu.Unlock()
+	return c.h
+}
+
+// TestCriticSoftTriggerIsMaxRuntime: the soft trigger handed to the host critic is
+// the run's resolved MaxRuntime (mort's two-tier model — the critic first wakes once
+// the run exceeds its nominal budget), NOT the global Defaults.CriticSoftTimeout.
+func TestCriticSoftTriggerIsMaxRuntime(t *testing.T) {
+	fp := fake.New("fake")
+	fp.Enqueue("m", fake.Reply("done"))
+	m, _ := fp.Model("m")
+	cc := &capturingCritic{h: &fakeCriticHandle{}}
+	ex := run.New(run.Config{
+		Registry: tool.NewRegistry(),
+		Models:   func(ctx context.Context, _ string) (context.Context, llm.Model, error) { return ctx, m, nil },
+		Ports:    run.Ports{Critic: cc},
+		Defaults: run.Defaults{CriticSoftTimeout: 90 * time.Second}, // distinct from MaxRuntime below
+	})
+	const wantSoft = 7 * time.Minute
+	ex.Run(context.Background(),
+		run.RunnableAgent{Name: "x", ModelTier: "m", MaxRuntime: wantSoft, Critic: run.CriticConfig{Enabled: true}},
+		tool.Invocation{RunID: "r"}, "go")
+	cc.mu.Lock()
+	got := cc.soft
+	cc.mu.Unlock()
+	if got != wantSoft {
+		t.Errorf("soft trigger = %v, want the agent's MaxRuntime %v (not Defaults.CriticSoftTimeout)", got, wantSoft)
+	}
+}

run/executor.go

@@ -30,6 +30,13 @@ type Defaults struct {
 	MaxSameToolCallRepeats   int           // retry-storm guard; default 3
 	CompactionThresholdRatio float64       // fraction of model context to compact at; default 0.7
 	CriticSoftTimeout        time.Duration // idle window before the critic wakes; default 90s
+	// CriticAbsoluteMax is the failsafe wall-clock ceiling for a critic-OWNED run
+	// (Ports.Critic set AND the agent enables it). For such a run MaxRuntime is the
+	// SOFT trigger, not a hard cap, and the critic's extendable backstop is the
+	// normal deadline — so this ceiling only fires if the critic never acts (a
+	// broken/nil host handle). Default 6h; never shorter than the run's MaxRuntime.
+	// Non-critic runs ignore it (they keep the literal MaxRuntime kill).
+	CriticAbsoluteMax time.Duration
 }
 
 func (d Defaults) withFallbacks() Defaults {
@@ -54,6 +61,9 @@ func (d Defaults) withFallbacks() Defaults {
 	if d.CriticSoftTimeout <= 0 {
 		d.CriticSoftTimeout = 90 * time.Second
 	}
+	if d.CriticAbsoluteMax <= 0 {
+		d.CriticAbsoluteMax = 6 * time.Hour
+	}
 	return d
 }
 
@@ -265,18 +275,36 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
 		postRun = st.PostRun
 	}
 
-	// Run context: bound by MaxRuntime, detached from the caller's deadline so a
+	// Run context: detached from the caller's deadline so a lane/queue wait doesn't
-	// lane/queue wait doesn't eat the run budget (mort's V10 lesson). Caller
+	// eat the run budget (mort's V10 lesson). Caller cancellation still propagates
-	// cancellation still propagates via MergeCancellation. Created BEFORE the
+	// via MergeCancellation. Created BEFORE the step observer so the observer
-	// step observer so the observer forwards the merged run context (not a
+	// forwards the merged run context (not a possibly-cancelled caller ctx) to
-	// possibly-cancelled caller ctx) to OnStep consumers.
+	// OnStep consumers.
-	// MaxRuntime stays a WithTimeout so its DeadlineExceeded propagates through the
+	//
-	// child chain (→ "timeout"), preserving the run's-own-timeout vs caller-cancel
+	// Two-tier timeout: who owns the hard deadline depends on the critic.
-	// distinction. A NESTED cause-carrying layer lets a critic kill surface as a
+	//   - NO critic (the default): MaxRuntime is a literal WithTimeout. Its
-	// distinct "killed" without disturbing that: only an ErrCriticKill cause is
+	//     DeadlineExceeded propagates through the child chain (→ "timeout"),
-	// consulted in statusFor; a generic run error or a caller cancel is classified
+	//     preserving the run's-own-timeout vs caller-cancel distinction.
-	// by the run error itself.
+	//   - critic OWNS the deadline (Ports.Critic set + ra.Critic.Enabled):
-	timeoutCtx, cancelTimeout := context.WithTimeout(context.WithoutCancel(ctx), maxRuntime)
+	//     MaxRuntime becomes the SOFT trigger (passed to startCritic), and the
+	//     critic's extendable backstop — watched in startCritic, which cancels via
+	//     cancelCause — is the real deadline. A slow-but-progressing run is given
+	//     room up to the backstop; only a stalled one is killed. We still wrap a
+	//     GENEROUS WithTimeout at CriticAbsoluteMax so a broken/nil critic handle
+	//     can't run unbounded; that ceiling never fires before the critic's backstop.
+	// A NESTED cause-carrying layer (cancelCause) lets a critic kill surface as a
+	// distinct "killed": only an ErrCriticKill cause is consulted in statusFor; a
+	// generic run error, a backstop expiry, or a caller cancel is classified by the
+	// run error itself.
+	criticOwnsDeadline := e.cfg.Ports.Critic != nil && ra.Critic.Enabled
+	hardCap := maxRuntime
+	if criticOwnsDeadline {
+		hardCap = e.cfg.Defaults.CriticAbsoluteMax
+		if hardCap < maxRuntime {
+			hardCap = maxRuntime // the failsafe ceiling is never shorter than the nominal budget
+		}
+	}
+	timeoutCtx, cancelTimeout := context.WithTimeout(context.WithoutCancel(ctx), hardCap)
 	defer cancelTimeout()
 	runCtx, cancelCause := context.WithCancelCause(timeoutCtx)
 	defer cancelCause(nil)
@@ -287,9 +315,11 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
 	checkpointCause = func() error { return context.Cause(runCtx) }
 
 	// Critic (optional): monitors the run for a stall, can nudge/extend/kill via
-	// its host Escalator. Its hard deadline is bound to runCtx (cancel on pass).
+	// its host Escalator. When it owns the deadline, MaxRuntime is its soft trigger
-	// nil-safe: no-op when no critic is configured or the agent doesn't enable it.
+	// (so a slow-but-progressing run survives past it); its extendable backstop is
-	critic, stopCritic := e.startCritic(runCtx, cancelCause, ra, info)
+	// bound to runCtx (cancel on pass). nil-safe: no-op when no critic is configured
+	// or the agent doesn't enable it.
+	critic, stopCritic := e.startCritic(runCtx, cancelCause, ra, info, maxRuntime)
 	defer stopCritic()
 
 	// Step instrumentation: accumulate Result.Steps + fire inv.OnStep, feed the