devrites 7.3.8 → 7.3.9

package/CHANGELOG.md

@@ -2,6 +2,12 @@
 
 All notable changes to DevRites are documented here. The format follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/) and DevRites adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). Releases are generated automatically by [semantic-release](https://semantic-release.gitbook.io/) from Conventional Commits on `main`.
 
+## [7.3.9](https://github.com/ViktorsBaikers/DevRites/compare/v7.3.8...v7.3.9) (2026-11-10)
+
+### Fixed
+
+* **devrites:** keep parallel autocomplete running past a source GO ([#212](https://github.com/ViktorsBaikers/DevRites/issues/212)) ([8dd7ab3](https://github.com/ViktorsBaikers/DevRites/commit/8dd7ab34a3d5ced36a7adb9ea4f2036db4116dcb))
+
 ## [7.3.8](https://github.com/ViktorsBaikers/DevRites/compare/v7.3.7...v7.3.8) (2026-11-09)
 
 ### Fixed

package/README.md

@@ -31,7 +31,7 @@ history can't lose it. After you clear the conversation, switch hosts, or hand
 the feature to another engineer, the next session reads the same plan,
 decisions, and evidence before it continues.
 
-**Status:** [`v7.3.8`](https://github.com/ViktorsBaikers/DevRites/releases/tag/v7.3.8): see [`CHANGELOG.md`](CHANGELOG.md) for release notes.
+**Status:** [`v7.3.9`](https://github.com/ViktorsBaikers/DevRites/releases/tag/v7.3.9): see [`CHANGELOG.md`](CHANGELOG.md) for release notes.
 
 This is the latest published release; `main` may contain unreleased work.
 

package/engine/internal/lib/continuation.go

@@ -58,7 +58,11 @@ func RunCheckContinuation(root string, args []string, stdout, stderr io.Writer)
 			return 2
 		}
 		if mode && frontierGlobalPriority(slug, frontier) || !mode && !verdict.Required && len(frontier) > 0 && frontier[0] == "frontier: actionable" {
-			verdict = continuationVerdict{Required: true, Invoke: "/rite-autocomplete", Source: slug, Reason: "global frontier requires native preparation, authority review, or selection; no Ship approval"}
+			reason := "global frontier requires native preparation, authority review, or selection; Git stays unauthorized"
+			if mode && frontierHasBatch(frontier) {
+				reason = "a source GO does not end autocomplete; run every launch command and every picks writer this turn; do not ask; Git stays unauthorized"
+			}
+			verdict = continuationVerdict{Required: true, Invoke: "/rite-autocomplete", Source: slug, Reason: reason}
 		} else if mode && frontierSourcePreparationWait(slug, frontier) {
 			verdict = continuationVerdict{Reason: frontierInheritedPreparationWait}
 		} else if mode && !verdict.Required {
@@ -67,13 +71,37 @@ func RunCheckContinuation(root string, args []string, stdout, stderr io.Writer)
 			}
 		}
 	}
-	writeContinuation(stdout, verdict)
+	writeContinuation(stdout, verdict, frontier)
 	for _, line := range frontier {
 		fmt.Fprintln(stdout, line)
 	}
 	return 0
 }
 
+// frontierHasBatch reports a capped same-turn batch: planning commands and/or
+// admitted slice writers. The controller invoke stays /rite-autocomplete.
+func frontierHasBatch(lines []string) bool {
+	for _, line := range lines {
+		if strings.HasPrefix(line, "frontier-concurrent: ") || strings.HasPrefix(line, "frontier-picks: ") {
+			return true
+		}
+	}
+	return false
+}
+
+func writeFrontierBatch(stdout io.Writer, lines []string) {
+	for _, line := range lines {
+		if cmd, ok := strings.CutPrefix(line, "frontier-concurrent: "); ok && cmd != "" {
+			fmt.Fprintf(stdout, "launch: %s\n", cmd)
+		}
+	}
+	for _, line := range lines {
+		if picks, ok := strings.CutPrefix(line, "frontier-picks: "); ok && picks != "" {
+			fmt.Fprintf(stdout, "picks: %s\n", picks)
+		}
+	}
+}
+
 func classifyContinuation(root, slug string) (continuationVerdict, error) {
 	featureDir, err := devritespaths.ExistingFeatureDirChecked(root, slug)
 	if err != nil {
@@ -246,11 +274,12 @@ func continuationArguments(command workflow.Command, args []string, slug string)
 	return badArgs()
 }
 
-func writeContinuation(stdout io.Writer, verdict continuationVerdict) {
+func writeContinuation(stdout io.Writer, verdict continuationVerdict, frontier []string) {
 	if verdict.Required {
 		fmt.Fprintln(stdout, "continuation: required")
 		fmt.Fprintf(stdout, "invoke: %s\n", verdict.Invoke)
 		fmt.Fprintf(stdout, "source: %s\n", verdict.Source)
+		writeFrontierBatch(stdout, frontier)
 	} else {
 		fmt.Fprintln(stdout, "continuation: clear")
 	}

package/engine/internal/lib/frontier.go

@@ -429,7 +429,7 @@ func FrontierLines(root, slug string) ([]string, error) {
 		}
 		if mode && action == "prepare" {
 			if cmd, ok := planningCursorCommand(reason); ok {
-				concurrent = append(concurrent, cmd)
+				concurrent = append(concurrent, commandWithWorkspace(cmd, name))
 			}
 		}
 		lines = append(lines, fmt.Sprintf("frontier-workspace: %s location=%s action=%s reason=%s", name, n.location, action, reason))
@@ -625,6 +625,20 @@ func agentOwnedPlanningCursor(n *frontierWorkspace) bool {
 	return preparationPhase(workflow.ForAction(n.next).Verb)
 }
 
+// commandWithWorkspace names the workspace on a phase command that omitted it.
+// `/rite-vet` cannot be launched; `/rite-vet <slug>` can.
+func commandWithWorkspace(cmd, slug string) string {
+	if slug == "" {
+		return cmd
+	}
+	for _, field := range strings.Fields(cmd) {
+		if field == slug {
+			return cmd
+		}
+	}
+	return cmd + " " + slug
+}
+
 func planningCursorCommand(reason string) (string, bool) {
 	const prefix = "native internal phase cursor: "
 	if !strings.HasPrefix(reason, prefix) {
@@ -740,10 +754,10 @@ func frontierLocal(root, controller string, n *frontierWorkspace, candidates *[]
 	if frontierSpecTarget(n) != "" {
 		return "complete", "native next Spec cursor recorded", nil
 	}
-	if _, internal, err := internalPhaseEdge(n.next, n.slug); err != nil {
+	if invoke, internal, err := internalPhaseEdge(n.next, n.slug); err != nil {
 		return "", "", err
 	} else if internal && (n.status == "running" || n.agentOwnedPlanning) {
-		return "prepare", "native internal phase cursor: " + n.next, nil
+		return "prepare", "native internal phase cursor: " + commandWithWorkspace(invoke, n.slug), nil
 	}
 	if command := workflow.ForAction(n.next); command.Verb == "ship" {
 		fields := strings.Fields(n.next)

package/engine/internal/lib/next.go

@@ -88,6 +88,7 @@ func RunNext(root string, args []string, stdout, stderr io.Writer) int {
 			return 2
 		}
 		if mode && frontierGlobalPriority(slug, lines) {
+			writeFrontierBatch(stdout, lines)
 			fmt.Fprintln(stdout, "next-command: /rite-autocomplete")
 			return exit
 		}

package/pack/.claude/skills/rite-autocomplete/reference/loop.md

@@ -44,9 +44,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/.claude/skills/rite-autocomplete/reference/stop-conditions.md

@@ -33,10 +33,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/claude/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/claude/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/codex/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/codex/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/devin/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/devin/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/dsh/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/dsh/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/omp/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](skill://rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/omp/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](skill://devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/pack/generated/pi/skills/rite-autocomplete/reference/loop.md

@@ -46,9 +46,9 @@ admission and the global pass below; local readiness is insufficient. Count ever
 Explicit activation resets only activation-local counters, not durable recovery.
 Pass resolved `--parallel N` to [Build](../../rite-build/reference/output.md#caller-owned-continuation).
 Legacy or invented execution ceilings never stop this controller; human limits bind.
-Resolve ranked validating/blocking picks via `devrites-engine state resolve`, not
-a user handoff. Escalating, irreversible-risk, access, unrecommended blocking picks
-and Seal judgment (`slice: seal`) still pause.
+Resolve ranked picks via `devrites-engine state resolve`, not
+a user handoff. Escalating, irreversible-risk, access, and unrecommended blocking
+picks pause that source only. A Seal GO does not end the run.
 
 ## Early preparation admission
 

package/pack/generated/pi/skills/rite-autocomplete/reference/stop-conditions.md

@@ -35,10 +35,10 @@ Autocomplete. Source human/Ship holds remain binding, while independent
 authorized planning can continue. `clear` is reply permission, not Ship approval.
 In autocomplete mode, `blocked` or `awaiting_human` is agent-owned recovery
 when no question cursor is set, no blocking or escalating question is open,
-and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. Ship, Build,
-a question cursor, and an open blocking or escalating question still wait.
-`frontier-concurrent:` lists at most the cap of those cursors. Before any reply,
-start every listed command and every `frontier-picks` writer in the same turn.
+and `next_action` is Spec, Clarify, Temper, Define, Plan, or Vet. A GO, Ship,
+Build, question, or blocking hold binds that source only and does not end the run.
+Before any reply, start every `frontier-concurrent:` command and every
+`frontier-picks` writer, at most the cap. Ask only when both are empty.
 
 A preparation refusal routes native work or a narrow disclosed scheduling
 deferral under the [global pass](../../devrites-lib/reference/parallel-workspaces.md#continuous-global-frontier).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "devrites",
-  "version": "7.3.8",
+  "version": "7.3.9",
   "description": "DevRites: a disciplined senior-engineer workflow pack for Claude Code, Codex, omp, pi, and Devin CLI",
   "license": "SEE LICENSE IN LICENSE",
   "homepage": "https://github.com/ViktorsBaikers/DevRites#readme",