oh-my-opencode 3.5.4 → 3.5.6

package/dist/index.js

@@ -4963,6 +4963,16 @@ Approach:
 - Draft with care
 - Polish for clarity and impact
 - Documentation, READMEs, articles, technical writing
+
+ANTI-AI-SLOP RULES (NON-NEGOTIABLE):
+- NEVER use em dashes (\u2014) or en dashes (\u2013). Use commas, periods, ellipses, or line breaks instead. Zero tolerance.
+- Remove AI-sounding phrases: "delve", "it's important to note", "I'd be happy to", "certainly", "please don't hesitate", "leverage", "utilize", "in order to", "moving forward", "circle back", "at the end of the day", "robust", "streamline", "facilitate"
+- Pick plain words. "Use" not "utilize". "Start" not "commence". "Help" not "facilitate".
+- Use contractions naturally: "don't" not "do not", "it's" not "it is".
+- Vary sentence length. Don't make every sentence the same length.
+- NEVER start consecutive sentences with the same word.
+- No filler openings: skip "In today's world...", "As we all know...", "It goes without saying..."
+- Write like a human, not a corporate template.
 </Category_Context>`, DEEP_CATEGORY_PROMPT_APPEND = `<Category_Context>
 You are working on GOAL-ORIENTED AUTONOMOUS tasks.
 
@@ -5238,14 +5248,14 @@ WHY THIS FORMAT IS MANDATORY:
 `, PLAN_AGENT_NAMES, PLAN_FAMILY_NAMES;
 var init_constants = __esm(() => {
   DEFAULT_CATEGORIES = {
-    "visual-engineering": { model: "google/gemini-3-pro" },
+    "visual-engineering": { model: "google/gemini-3-pro", variant: "high" },
     ultrabrain: { model: "openai/gpt-5.3-codex", variant: "xhigh" },
     deep: { model: "openai/gpt-5.3-codex", variant: "medium" },
     artistry: { model: "google/gemini-3-pro", variant: "high" },
     quick: { model: "anthropic/claude-haiku-4-5" },
     "unspecified-low": { model: "anthropic/claude-sonnet-4-5" },
     "unspecified-high": { model: "anthropic/claude-opus-4-6", variant: "max" },
-    writing: { model: "google/gemini-3-flash" }
+    writing: { model: "kimi-for-coding/k2p5" }
   };
   CATEGORY_PROMPT_APPENDS = {
     "visual-engineering": VISUAL_CATEGORY_PROMPT_APPEND,
@@ -12230,6 +12240,8 @@ var TOAST_DURATION_MS = 900;
 var COUNTDOWN_GRACE_PERIOD_MS = 500;
 var ABORT_WINDOW_MS = 3000;
 var CONTINUATION_COOLDOWN_MS = 30000;
+var MAX_CONSECUTIVE_FAILURES = 5;
+var FAILURE_RESET_WINDOW_MS = 5 * 60 * 1000;
 
 // src/hooks/todo-continuation-enforcer/handler.ts
 init_logger();
@@ -12454,11 +12466,14 @@ ${todoList}`;
     if (injectionState) {
       injectionState.inFlight = false;
       injectionState.lastInjectedAt = Date.now();
+      injectionState.consecutiveFailures = 0;
     }
   } catch (error) {
     log(`[${HOOK_NAME}] Injection failed`, { sessionID, error: String(error) });
     if (injectionState) {
       injectionState.inFlight = false;
+      injectionState.lastInjectedAt = Date.now();
+      injectionState.consecutiveFailures = (injectionState.consecutiveFailures ?? 0) + 1;
     }
   }
 }
@@ -12577,8 +12592,28 @@ async function handleSessionIdle(args) {
     log(`[${HOOK_NAME}] Skipped: injection in flight`, { sessionID });
     return;
   }
-  if (state.lastInjectedAt && Date.now() - state.lastInjectedAt < CONTINUATION_COOLDOWN_MS) {
-    log(`[${HOOK_NAME}] Skipped: cooldown active`, { sessionID });
+  if (state.consecutiveFailures >= MAX_CONSECUTIVE_FAILURES && state.lastInjectedAt && Date.now() - state.lastInjectedAt >= FAILURE_RESET_WINDOW_MS) {
+    state.consecutiveFailures = 0;
+    log(`[${HOOK_NAME}] Reset consecutive failures after recovery window`, {
+      sessionID,
+      failureResetWindowMs: FAILURE_RESET_WINDOW_MS
+    });
+  }
+  if (state.consecutiveFailures >= MAX_CONSECUTIVE_FAILURES) {
+    log(`[${HOOK_NAME}] Skipped: max consecutive failures reached`, {
+      sessionID,
+      consecutiveFailures: state.consecutiveFailures,
+      maxConsecutiveFailures: MAX_CONSECUTIVE_FAILURES
+    });
+    return;
+  }
+  const effectiveCooldown = CONTINUATION_COOLDOWN_MS * Math.pow(2, Math.min(state.consecutiveFailures, 5));
+  if (state.lastInjectedAt && Date.now() - state.lastInjectedAt < effectiveCooldown) {
+    log(`[${HOOK_NAME}] Skipped: cooldown active`, {
+      sessionID,
+      effectiveCooldown,
+      consecutiveFailures: state.consecutiveFailures
+    });
     return;
   }
   let resolvedInfo;
@@ -12767,7 +12802,9 @@ function createSessionStateStore() {
       existing.lastAccessedAt = Date.now();
       return existing.state;
     }
-    const state = {};
+    const state = {
+      consecutiveFailures: 0
+    };
     sessions.set(sessionID, { state, lastAccessedAt: Date.now() });
     return state;
   }
@@ -28285,11 +28322,17 @@ function createRecoveryState() {
 function createAnthropicContextWindowLimitRecoveryHook(ctx, options) {
   const autoCompactState = createRecoveryState();
   const experimental = options?.experimental;
+  const pendingCompactionTimeoutBySession = new Map;
   const eventHandler = async ({ event }) => {
     const props = event.properties;
     if (event.type === "session.deleted") {
       const sessionInfo = props?.info;
       if (sessionInfo?.id) {
+        const timeoutID = pendingCompactionTimeoutBySession.get(sessionInfo.id);
+        if (timeoutID !== undefined) {
+          clearTimeout(timeoutID);
+          pendingCompactionTimeoutBySession.delete(sessionInfo.id);
+        }
         autoCompactState.pendingCompact.delete(sessionInfo.id);
         autoCompactState.errorDataBySession.delete(sessionInfo.id);
         autoCompactState.retryStateBySession.delete(sessionInfo.id);
@@ -28324,9 +28367,11 @@ function createAnthropicContextWindowLimitRecoveryHook(ctx, options) {
             duration: 3000
           }
         }).catch(() => {});
-        setTimeout(() => {
+        const timeoutID = setTimeout(() => {
+          pendingCompactionTimeoutBySession.delete(sessionID);
           executeCompact(sessionID, { providerID, modelID }, autoCompactState, ctx.client, ctx.directory, experimental);
         }, 300);
+        pendingCompactionTimeoutBySession.set(sessionID, timeoutID);
       }
       return;
     }
@@ -28352,6 +28397,11 @@ function createAnthropicContextWindowLimitRecoveryHook(ctx, options) {
         return;
       if (!autoCompactState.pendingCompact.has(sessionID))
         return;
+      const timeoutID = pendingCompactionTimeoutBySession.get(sessionID);
+      if (timeoutID !== undefined) {
+        clearTimeout(timeoutID);
+        pendingCompactionTimeoutBySession.delete(sessionID);
+      }
       const errorData = autoCompactState.errorDataBySession.get(sessionID);
       const lastAssistant = await getLastAssistant(sessionID, ctx.client, ctx.directory);
       if (lastAssistant?.summary === true) {
@@ -33130,9 +33180,10 @@ var AGENT_MODEL_REQUIREMENTS = {
 var CATEGORY_MODEL_REQUIREMENTS = {
   "visual-engineering": {
     fallbackChain: [
-      { providers: ["google", "github-copilot", "opencode"], model: "gemini-3-pro" },
+      { providers: ["google", "github-copilot", "opencode"], model: "gemini-3-pro", variant: "high" },
+      { providers: ["zai-coding-plan"], model: "glm-5" },
       { providers: ["anthropic", "github-copilot", "opencode"], model: "claude-opus-4-6", variant: "max" },
-      { providers: ["zai-coding-plan"], model: "glm-4.7" }
+      { providers: ["kimi-for-coding"], model: "k2p5" }
     ]
   },
   ultrabrain: {
@@ -33181,10 +33232,9 @@ var CATEGORY_MODEL_REQUIREMENTS = {
   },
   writing: {
     fallbackChain: [
+      { providers: ["kimi-for-coding"], model: "k2p5" },
       { providers: ["google", "github-copilot", "opencode"], model: "gemini-3-flash" },
-      { providers: ["anthropic", "github-copilot", "opencode"], model: "claude-sonnet-4-5" },
-      { providers: ["zai-coding-plan"], model: "glm-4.7" },
-      { providers: ["openai", "github-copilot", "opencode"], model: "gpt-5.2" }
+      { providers: ["anthropic", "github-copilot", "opencode"], model: "claude-sonnet-4-5" }
     ]
   }
 };
@@ -44976,6 +45026,7 @@ function createUnstableAgentBabysitterHook(ctx, options) {
   };
 }
 // src/hooks/preemptive-compaction.ts
+init_logger();
 var DEFAULT_ACTUAL_LIMIT = 200000;
 var ANTHROPIC_ACTUAL_LIMIT3 = process.env.ANTHROPIC_1M_CONTEXT === "true" || process.env.VERTEX_ANTHROPIC_1M_CONTEXT === "true" ? 1e6 : DEFAULT_ACTUAL_LIMIT;
 var PREEMPTIVE_COMPACTION_THRESHOLD = 0.78;
@@ -45007,7 +45058,9 @@ function createPreemptiveCompactionHook(ctx) {
         query: { directory: ctx.directory }
       });
       compactedSessions.add(sessionID);
-    } catch {} finally {
+    } catch (error45) {
+      log("[preemptive-compaction] Compaction failed", { sessionID, error: String(error45) });
+    } finally {
       compactionInProgress.delete(sessionID);
     }
   };
@@ -51135,6 +51188,7 @@ Task ID: ${task.id}`;
     const pollStart = Date.now();
     let lastMsgCount = 0;
     let stablePolls = 0;
+    let terminalStatus;
     while (Date.now() - pollStart < timingCfg.MAX_POLL_TIME_MS) {
       if (ctx.abort?.aborted) {
         return `Task aborted (was running in background mode).
@@ -51142,6 +51196,11 @@ Task ID: ${task.id}`;
 Session ID: ${sessionID}`;
       }
       await new Promise((resolve10) => setTimeout(resolve10, timingCfg.POLL_INTERVAL_MS));
+      const currentTask = manager.getTask(task.id);
+      if (currentTask && (currentTask.status === "interrupt" || currentTask.status === "error" || currentTask.status === "cancelled")) {
+        terminalStatus = { status: currentTask.status, error: currentTask.error };
+        break;
+      }
       const statusResult = await client2.session.status();
       const allStatuses = statusResult.data ?? {};
       const sessionStatus = allStatuses[sessionID];
@@ -51164,6 +51223,23 @@ Session ID: ${sessionID}`;
         lastMsgCount = currentMsgCount;
       }
     }
+    if (terminalStatus) {
+      const duration4 = formatDuration2(startTime);
+      return `SUPERVISED TASK FAILED (${terminalStatus.status})
+
+Task was interrupted/failed while running in monitored background mode.
+${terminalStatus.error ? `Error: ${terminalStatus.error}` : ""}
+
+Duration: ${duration4}
+Agent: ${agentToUse}${args.category ? ` (category: ${args.category})` : ""}
+Model: ${actualModel}
+
+The task session may contain partial results.
+
+<task_metadata>
+session_id: ${sessionID}
+</task_metadata>`;
+    }
     const messagesResult = await client2.session.messages({ path: { id: sessionID } });
     const messages = messagesResult.data ?? messagesResult;
     const assistantMessages = messages.filter((m) => m.info?.role === "assistant").sort((a, b) => (b.info?.time?.created ?? 0) - (a.info?.time?.created ?? 0));
@@ -53497,6 +53573,11 @@ class BackgroundManager {
       });
       return existingTask;
     }
+    const completionTimer = this.completionTimers.get(existingTask.id);
+    if (completionTimer) {
+      clearTimeout(completionTimer);
+      this.completionTimers.delete(existingTask.id);
+    }
     const concurrencyKey = existingTask.concurrencyGroup ?? existingTask.agent;
     await this.concurrencyManager.acquire(concurrencyKey);
     existingTask.concurrencyKey = concurrencyKey;
@@ -53600,7 +53681,7 @@ class BackgroundManager {
   }
   handleEvent(event) {
     const props = event.properties;
-    if (event.type === "message.part.updated") {
+    if (event.type === "message.part.updated" || event.type === "message.part.delta") {
       if (!props || typeof props !== "object" || !("sessionID" in props))
         return;
       const partInfo = props;
@@ -53704,6 +53785,10 @@ class BackgroundManager {
       this.cleanupPendingByParent(task);
       this.tasks.delete(task.id);
       this.clearNotificationsForTask(task.id);
+      const toastManager = getTaskToastManager();
+      if (toastManager) {
+        toastManager.removeTask(task.id);
+      }
       if (task.sessionID) {
         subagentSessions.delete(task.sessionID);
       }
@@ -53746,6 +53831,10 @@ class BackgroundManager {
         this.cleanupPendingByParent(task);
         this.tasks.delete(task.id);
         this.clearNotificationsForTask(task.id);
+        const toastManager = getTaskToastManager();
+        if (toastManager) {
+          toastManager.removeTask(task.id);
+        }
         if (task.sessionID) {
           subagentSessions.delete(task.sessionID);
         }
@@ -53860,6 +53949,10 @@ class BackgroundManager {
       }).catch(() => {});
     }
     if (options?.skipNotification) {
+      const toastManager = getTaskToastManager();
+      if (toastManager) {
+        toastManager.removeTask(task.id);
+      }
       log(`[background-agent] Task cancelled via ${source} (notification skipped):`, task.id);
       return true;
     }
@@ -54037,11 +54130,10 @@ Use \`background_output(task_id="${task.id}")\` to retrieve this result when rea
       }
     } catch (error45) {
       if (this.isAbortedSessionError(error45)) {
-        log("[background-agent] Parent session aborted, skipping notification:", {
+        log("[background-agent] Parent session aborted while loading messages; using messageDir fallback:", {
           taskId: task.id,
           parentSessionID: task.parentSessionID
         });
-        return;
       }
       const messageDir = getMessageDir12(task.parentSessionID);
       const currentMessage = messageDir ? findNearestMessageWithFields(messageDir) : null;
@@ -54071,13 +54163,13 @@ Use \`background_output(task_id="${task.id}")\` to retrieve this result when rea
       });
     } catch (error45) {
       if (this.isAbortedSessionError(error45)) {
-        log("[background-agent] Parent session aborted, skipping notification:", {
+        log("[background-agent] Parent session aborted while sending notification; continuing cleanup:", {
           taskId: task.id,
           parentSessionID: task.parentSessionID
         });
-        return;
+      } else {
+        log("[background-agent] Failed to send notification:", error45);
       }
-      log("[background-agent] Failed to send notification:", error45);
     }
     if (allComplete) {
       for (const completedTask of completedTasks) {
@@ -54190,6 +54282,10 @@ Use \`background_output(task_id="${task.id}")\` to retrieve this result when rea
           }
         }
         this.clearNotificationsForTask(taskId);
+        const toastManager = getTaskToastManager();
+        if (toastManager) {
+          toastManager.removeTask(taskId);
+        }
         this.tasks.delete(taskId);
         if (task.sessionID) {
           subagentSessions.delete(task.sessionID);
@@ -54225,7 +54321,8 @@ Use \`background_output(task_id="${task.id}")\` to retrieve this result when rea
       const sessionID = task.sessionID;
       if (!startedAt || !sessionID)
         continue;
-      const sessionIsRunning = allStatuses[sessionID]?.type === "running";
+      const sessionStatus = allStatuses[sessionID]?.type;
+      const sessionIsRunning = sessionStatus !== undefined && sessionStatus !== "idle";
       const runtime = now - startedAt.getTime();
       if (!task.progress?.lastUpdate) {
         if (sessionIsRunning)
@@ -64430,7 +64527,21 @@ Your ONLY valid output locations are \`.sisyphus/plans/*.md\` and \`.sisyphus/dr
 
 Example: \`.sisyphus/plans/auth-refactor.md\`
 
-### 5. SINGLE PLAN MANDATE (CRITICAL)
+### 5. MAXIMUM PARALLELISM PRINCIPLE (NON-NEGOTIABLE)
+
+Your plans MUST maximize parallel execution. This is a core planning quality metric.
+
+**Granularity Rule**: One task = one module/concern = 1-3 files.
+If a task touches 4+ files or 2+ unrelated concerns, SPLIT IT.
+
+**Parallelism Target**: Aim for 5-8 tasks per wave.
+If any wave has fewer than 3 tasks (except the final integration), you under-split.
+
+**Dependency Minimization**: Structure tasks so shared dependencies
+(types, interfaces, configs) are extracted as early Wave-1 tasks,
+unblocking maximum parallelism in subsequent waves.
+
+### 6. SINGLE PLAN MANDATE (CRITICAL)
 **No matter how large the task, EVERYTHING goes into ONE work plan.**
 
 **NEVER:**
@@ -64453,7 +64564,7 @@ Example: \`.sisyphus/plans/auth-refactor.md\`
 
 **The plan can have 50+ TODOs. That's OK. ONE PLAN.**
 
-### 5.1 SINGLE ATOMIC WRITE (CRITICAL - Prevents Content Loss)
+### 6.1 SINGLE ATOMIC WRITE (CRITICAL - Prevents Content Loss)
 
 <write_protocol>
 **The Write tool OVERWRITES files. It does NOT append.**
@@ -64496,7 +64607,7 @@ Example: \`.sisyphus/plans/auth-refactor.md\`
 - [ ] File already exists with my content? \u2192 Use Edit to append, NOT Write
 </write_protocol>
 
-### 6. DRAFT AS WORKING MEMORY (MANDATORY)
+### 7. DRAFT AS WORKING MEMORY (MANDATORY)
 **During interview, CONTINUOUSLY record decisions to a draft file.**
 
 **Draft Location**: \`.sisyphus/drafts/{name}.md\`
@@ -65303,108 +65414,25 @@ Generate plan to: \`.sisyphus/plans/{name}.md\`
 
 ## Verification Strategy (MANDATORY)
 
-> **UNIVERSAL RULE: ZERO HUMAN INTERVENTION**
->
-> ALL tasks in this plan MUST be verifiable WITHOUT any human action.
-> This is NOT conditional \u2014 it applies to EVERY task, regardless of test strategy.
->
-> **FORBIDDEN** \u2014 acceptance criteria that require:
-> - "User manually tests..." / "\uC0AC\uC6A9\uC790\uAC00 \uC9C1\uC811 \uD14C\uC2A4\uD2B8..."
-> - "User visually confirms..." / "\uC0AC\uC6A9\uC790\uAC00 \uB208\uC73C\uB85C \uD655\uC778..."
-> - "User interacts with..." / "\uC0AC\uC6A9\uC790\uAC00 \uC9C1\uC811 \uC870\uC791..."
-> - "Ask user to verify..." / "\uC0AC\uC6A9\uC790\uC5D0\uAC8C \uD655\uC778 \uC694\uCCAD..."
-> - ANY step where a human must perform an action
->
-> **ALL verification is executed by the agent** using tools (Playwright, interactive_bash, curl, etc.). No exceptions.
+> **ZERO HUMAN INTERVENTION** \u2014 ALL verification is agent-executed. No exceptions.
+> Acceptance criteria requiring "user manually tests/confirms" are FORBIDDEN.
 
 ### Test Decision
 - **Infrastructure exists**: [YES/NO]
 - **Automated tests**: [TDD / Tests-after / None]
 - **Framework**: [bun test / vitest / jest / pytest / none]
+- **If TDD**: Each task follows RED (failing test) \u2192 GREEN (minimal impl) \u2192 REFACTOR
 
-### If TDD Enabled
-
-Each TODO follows RED-GREEN-REFACTOR:
-
-**Task Structure:**
-1. **RED**: Write failing test first
-   - Test file: \`[path].test.ts\`
-   - Test command: \`bun test [file]\`
-   - Expected: FAIL (test exists, implementation doesn't)
-2. **GREEN**: Implement minimum code to pass
-   - Command: \`bun test [file]\`
-   - Expected: PASS
-3. **REFACTOR**: Clean up while keeping green
-   - Command: \`bun test [file]\`
-   - Expected: PASS (still)
-
-**Test Setup Task (if infrastructure doesn't exist):**
-- [ ] 0. Setup Test Infrastructure
-  - Install: \`bun add -d [test-framework]\`
-  - Config: Create \`[config-file]\`
-  - Verify: \`bun test --help\` \u2192 shows help
-  - Example: Create \`src/__tests__/example.test.ts\`
-  - Verify: \`bun test\` \u2192 1 test passes
-
-### Agent-Executed QA Scenarios (MANDATORY \u2014 ALL tasks)
-
-> Whether TDD is enabled or not, EVERY task MUST include Agent-Executed QA Scenarios.
-> - **With TDD**: QA scenarios complement unit tests at integration/E2E level
-> - **Without TDD**: QA scenarios are the PRIMARY verification method
->
-> These describe how the executing agent DIRECTLY verifies the deliverable
-> by running it \u2014 opening browsers, executing commands, sending API requests.
-> The agent performs what a human tester would do, but automated via tools.
-
-**Verification Tool by Deliverable Type:**
-
-| Type | Tool | How Agent Verifies |
-|------|------|-------------------|
-| **Frontend/UI** | Playwright (playwright skill) | Navigate, interact, assert DOM, screenshot |
-| **TUI/CLI** | interactive_bash (tmux) | Run command, send keystrokes, validate output |
-| **API/Backend** | Bash (curl/httpie) | Send requests, parse responses, assert fields |
-| **Library/Module** | Bash (bun/node REPL) | Import, call functions, compare output |
-| **Config/Infra** | Bash (shell commands) | Apply config, run state checks, validate |
-
-**Each Scenario MUST Follow This Format:**
-
-\`\`\`
-Scenario: [Descriptive name \u2014 what user action/flow is being verified]
-  Tool: [Playwright / interactive_bash / Bash]
-  Preconditions: [What must be true before this scenario runs]
-  Steps:
-    1. [Exact action with specific selector/command/endpoint]
-    2. [Next action with expected intermediate state]
-    3. [Assertion with exact expected value]
-  Expected Result: [Concrete, observable outcome]
-  Failure Indicators: [What would indicate failure]
-  Evidence: [Screenshot path / output capture / response body path]
-\`\`\`
+### QA Policy
+Every task MUST include agent-executed QA scenarios (see TODO template below).
+Evidence saved to \`.sisyphus/evidence/task-{N}-{scenario-slug}.{ext}\`.
 
-**Scenario Detail Requirements:**
-- **Selectors**: Specific CSS selectors (\`.login-button\`, not "the login button")
-- **Data**: Concrete test data (\`"test@example.com"\`, not \`"[email]"\`)
-- **Assertions**: Exact values (\`text contains "Welcome back"\`, not "verify it works")
-- **Timing**: Include wait conditions where relevant (\`Wait for .dashboard (timeout: 10s)\`)
-- **Negative Scenarios**: At least ONE failure/error scenario per feature
-- **Evidence Paths**: Specific file paths (\`.sisyphus/evidence/task-N-scenario-name.png\`)
-
-**Anti-patterns (NEVER write scenarios like this):**
-- \u274C "Verify the login page works correctly"
-- \u274C "Check that the API returns the right data"
-- \u274C "Test the form validation"
-- \u274C "User opens browser and confirms..."
-
-**Write scenarios like this instead:**
-- \u2705 \`Navigate to /login \u2192 Fill input[name="email"] with "test@example.com" \u2192 Fill input[name="password"] with "Pass123!" \u2192 Click button[type="submit"] \u2192 Wait for /dashboard \u2192 Assert h1 contains "Welcome"\`
-- \u2705 \`POST /api/users {"name":"Test","email":"new@test.com"} \u2192 Assert status 201 \u2192 Assert response.id is UUID \u2192 GET /api/users/{id} \u2192 Assert name equals "Test"\`
-- \u2705 \`Run ./cli --config test.yaml \u2192 Wait for "Loaded" in stdout \u2192 Send "q" \u2192 Assert exit code 0 \u2192 Assert stdout contains "Goodbye"\`
-
-**Evidence Requirements:**
-- Screenshots: \`.sisyphus/evidence/\` for all UI verifications
-- Terminal output: Captured for CLI/TUI verifications
-- Response bodies: Saved for API verifications
-- All evidence referenced by specific file path in acceptance criteria
+| Deliverable Type | Verification Tool | Method |
+|------------------|-------------------|--------|
+| Frontend/UI | Playwright (playwright skill) | Navigate, interact, assert DOM, screenshot |
+| TUI/CLI | interactive_bash (tmux) | Run command, send keystrokes, validate output |
+| API/Backend | Bash (curl) | Send requests, assert status + response fields |
+| Library/Module | Bash (bun/node REPL) | Import, call functions, compare output |
 
 ---
 
@@ -65414,49 +65442,82 @@ Scenario: [Descriptive name \u2014 what user action/flow is being verified]
 
 > Maximize throughput by grouping independent tasks into parallel waves.
 > Each wave completes before the next begins.
+> Target: 5-8 tasks per wave. Fewer than 3 per wave (except final) = under-splitting.
 
 \`\`\`
-Wave 1 (Start Immediately):
-\u251C\u2500\u2500 Task 1: [no dependencies]
-\u2514\u2500\u2500 Task 5: [no dependencies]
-
-Wave 2 (After Wave 1):
-\u251C\u2500\u2500 Task 2: [depends: 1]
-\u251C\u2500\u2500 Task 3: [depends: 1]
-\u2514\u2500\u2500 Task 6: [depends: 5]
-
-Wave 3 (After Wave 2):
-\u2514\u2500\u2500 Task 4: [depends: 2, 3]
-
-Critical Path: Task 1 \u2192 Task 2 \u2192 Task 4
-Parallel Speedup: ~40% faster than sequential
+Wave 1 (Start Immediately \u2014 foundation + scaffolding):
+\u251C\u2500\u2500 Task 1: Project scaffolding + config [quick]
+\u251C\u2500\u2500 Task 2: Design system tokens [quick]
+\u251C\u2500\u2500 Task 3: Type definitions [quick]
+\u251C\u2500\u2500 Task 4: Schema definitions [quick]
+\u251C\u2500\u2500 Task 5: Storage interface + in-memory impl [quick]
+\u251C\u2500\u2500 Task 6: Auth middleware [quick]
+\u2514\u2500\u2500 Task 7: Client module [quick]
+
+Wave 2 (After Wave 1 \u2014 core modules, MAX PARALLEL):
+\u251C\u2500\u2500 Task 8: Core business logic (depends: 3, 5, 7) [deep]
+\u251C\u2500\u2500 Task 9: API endpoints (depends: 4, 5) [unspecified-high]
+\u251C\u2500\u2500 Task 10: Secondary storage impl (depends: 5) [unspecified-high]
+\u251C\u2500\u2500 Task 11: Retry/fallback logic (depends: 8) [deep]
+\u251C\u2500\u2500 Task 12: UI layout + navigation (depends: 2) [visual-engineering]
+\u251C\u2500\u2500 Task 13: API client + hooks (depends: 4) [quick]
+\u2514\u2500\u2500 Task 14: Telemetry middleware (depends: 5, 10) [unspecified-high]
+
+Wave 3 (After Wave 2 \u2014 integration + UI):
+\u251C\u2500\u2500 Task 15: Main route combining modules (depends: 6, 11, 14) [deep]
+\u251C\u2500\u2500 Task 16: UI data visualization (depends: 12, 13) [visual-engineering]
+\u251C\u2500\u2500 Task 17: Deployment config A (depends: 15) [quick]
+\u251C\u2500\u2500 Task 18: Deployment config B (depends: 15) [quick]
+\u251C\u2500\u2500 Task 19: Deployment config C (depends: 15) [quick]
+\u2514\u2500\u2500 Task 20: UI request log + build (depends: 16) [visual-engineering]
+
+Wave 4 (After Wave 3 \u2014 verification):
+\u251C\u2500\u2500 Task 21: Integration tests (depends: 15) [deep]
+\u251C\u2500\u2500 Task 22: UI QA - Playwright (depends: 20) [unspecified-high]
+\u251C\u2500\u2500 Task 23: E2E QA (depends: 21) [deep]
+\u2514\u2500\u2500 Task 24: Git cleanup + tagging (depends: 21) [git]
+
+Wave FINAL (After ALL tasks \u2014 independent review, 4 parallel):
+\u251C\u2500\u2500 Task F1: Plan compliance audit (oracle)
+\u251C\u2500\u2500 Task F2: Code quality review (unspecified-high)
+\u251C\u2500\u2500 Task F3: Real manual QA (unspecified-high)
+\u2514\u2500\u2500 Task F4: Scope fidelity check (deep)
+
+Critical Path: Task 1 \u2192 Task 5 \u2192 Task 8 \u2192 Task 11 \u2192 Task 15 \u2192 Task 21 \u2192 F1-F4
+Parallel Speedup: ~70% faster than sequential
+Max Concurrent: 7 (Waves 1 & 2)
 \`\`\`
 
-### Dependency Matrix
+### Dependency Matrix (abbreviated \u2014 show ALL tasks in your generated plan)
 
-| Task | Depends On | Blocks | Can Parallelize With |
-|------|------------|--------|---------------------|
-| 1 | None | 2, 3 | 5 |
-| 2 | 1 | 4 | 3, 6 |
-| 3 | 1 | 4 | 2, 6 |
-| 4 | 2, 3 | None | None (final) |
-| 5 | None | 6 | 1 |
-| 6 | 5 | None | 2, 3 |
+| Task | Depends On | Blocks | Wave |
+|------|------------|--------|------|
+| 1-7 | \u2014 | 8-14 | 1 |
+| 8 | 3, 5, 7 | 11, 15 | 2 |
+| 11 | 8 | 15 | 2 |
+| 14 | 5, 10 | 15 | 2 |
+| 15 | 6, 11, 14 | 17-19, 21 | 3 |
+| 21 | 15 | 23, 24 | 4 |
+
+> This is abbreviated for reference. YOUR generated plan must include the FULL matrix for ALL tasks.
 
 ### Agent Dispatch Summary
 
-| Wave | Tasks | Recommended Agents |
-|------|-------|-------------------|
-| 1 | 1, 5 | task(category="...", load_skills=[...], run_in_background=false) |
-| 2 | 2, 3, 6 | dispatch parallel after Wave 1 completes |
-| 3 | 4 | final integration task |
+| Wave | # Parallel | Tasks \u2192 Agent Category |
+|------|------------|----------------------|
+| 1 | **7** | T1-T4 \u2192 \`quick\`, T5 \u2192 \`quick\`, T6 \u2192 \`quick\`, T7 \u2192 \`quick\` |
+| 2 | **7** | T8 \u2192 \`deep\`, T9 \u2192 \`unspecified-high\`, T10 \u2192 \`unspecified-high\`, T11 \u2192 \`deep\`, T12 \u2192 \`visual-engineering\`, T13 \u2192 \`quick\`, T14 \u2192 \`unspecified-high\` |
+| 3 | **6** | T15 \u2192 \`deep\`, T16 \u2192 \`visual-engineering\`, T17-T19 \u2192 \`quick\`, T20 \u2192 \`visual-engineering\` |
+| 4 | **4** | T21 \u2192 \`deep\`, T22 \u2192 \`unspecified-high\`, T23 \u2192 \`deep\`, T24 \u2192 \`git\` |
+| FINAL | **4** | F1 \u2192 \`oracle\`, F2 \u2192 \`unspecified-high\`, F3 \u2192 \`unspecified-high\`, F4 \u2192 \`deep\` |
 
 ---
 
 ## TODOs
 
 > Implementation + Test = ONE Task. Never separate.
-> EVERY task MUST have: Recommended Agent Profile + Parallelization info.
+> EVERY task MUST have: Recommended Agent Profile + Parallelization info + QA Scenarios.
+> **A task WITHOUT QA Scenarios is INCOMPLETE. No exceptions.**
 
 - [ ] 1. [Task Title]
 
@@ -65490,22 +65551,15 @@ Parallel Speedup: ~40% faster than sequential
 
   **Pattern References** (existing code to follow):
   - \`src/services/auth.ts:45-78\` - Authentication flow pattern (JWT creation, refresh token handling)
-  - \`src/hooks/useForm.ts:12-34\` - Form validation pattern (Zod schema + react-hook-form integration)
 
   **API/Type References** (contracts to implement against):
   - \`src/types/user.ts:UserDTO\` - Response shape for user endpoints
-  - \`src/api/schema.ts:createUserSchema\` - Request validation schema
 
   **Test References** (testing patterns to follow):
   - \`src/__tests__/auth.test.ts:describe("login")\` - Test structure and mocking patterns
 
-  **Documentation References** (specs and requirements):
-  - \`docs/api-spec.md#authentication\` - API contract details
-  - \`ARCHITECTURE.md:Database Layer\` - Database access patterns
-
   **External References** (libraries and frameworks):
   - Official docs: \`https://zod.dev/?id=basic-usage\` - Zod validation syntax
-  - Example repo: \`github.com/example/project/src/auth\` - Reference implementation
 
   **WHY Each Reference Matters** (explain the relevance):
   - Don't just list files - explain what pattern/information the executor should extract
@@ -65516,113 +65570,60 @@ Parallel Speedup: ~40% faster than sequential
 
   > **AGENT-EXECUTABLE VERIFICATION ONLY** \u2014 No human action permitted.
   > Every criterion MUST be verifiable by running a command or using a tool.
-  > REPLACE all placeholders with actual values from task context.
 
   **If TDD (tests enabled):**
   - [ ] Test file created: src/auth/login.test.ts
-  - [ ] Test covers: successful login returns JWT token
   - [ ] bun test src/auth/login.test.ts \u2192 PASS (3 tests, 0 failures)
 
-  **Agent-Executed QA Scenarios (MANDATORY \u2014 per-scenario, ultra-detailed):**
-
-  > Write MULTIPLE named scenarios per task: happy path AND failure cases.
-  > Each scenario = exact tool + steps with real selectors/data + evidence path.
-
-  **Example \u2014 Frontend/UI (Playwright):**
+  **QA Scenarios (MANDATORY \u2014 task is INCOMPLETE without these):**
 
-  \\\`\\\`\\\`
-  Scenario: Successful login redirects to dashboard
-    Tool: Playwright (playwright skill)
-    Preconditions: Dev server running on localhost:3000, test user exists
-    Steps:
-      1. Navigate to: http://localhost:3000/login
-      2. Wait for: input[name="email"] visible (timeout: 5s)
-      3. Fill: input[name="email"] \u2192 "test@example.com"
-      4. Fill: input[name="password"] \u2192 "ValidPass123!"
-      5. Click: button[type="submit"]
-      6. Wait for: navigation to /dashboard (timeout: 10s)
-      7. Assert: h1 text contains "Welcome back"
-      8. Assert: cookie "session_token" exists
-      9. Screenshot: .sisyphus/evidence/task-1-login-success.png
-    Expected Result: Dashboard loads with welcome message
-    Evidence: .sisyphus/evidence/task-1-login-success.png
-
-  Scenario: Login fails with invalid credentials
-    Tool: Playwright (playwright skill)
-    Preconditions: Dev server running, no valid user with these credentials
-    Steps:
-      1. Navigate to: http://localhost:3000/login
-      2. Fill: input[name="email"] \u2192 "wrong@example.com"
-      3. Fill: input[name="password"] \u2192 "WrongPass"
-      4. Click: button[type="submit"]
-      5. Wait for: .error-message visible (timeout: 5s)
-      6. Assert: .error-message text contains "Invalid credentials"
-      7. Assert: URL is still /login (no redirect)
-      8. Screenshot: .sisyphus/evidence/task-1-login-failure.png
-    Expected Result: Error message shown, stays on login page
-    Evidence: .sisyphus/evidence/task-1-login-failure.png
-  \\\`\\\`\\\`
-
-  **Example \u2014 API/Backend (curl):**
+  > **This is NOT optional. A task without QA scenarios WILL BE REJECTED.**
+  >
+  > Write scenario tests that verify the ACTUAL BEHAVIOR of what you built.
+  > Minimum: 1 happy path + 1 failure/edge case per task.
+  > Each scenario = exact tool + exact steps + exact assertions + evidence path.
+  >
+  > **The executing agent MUST run these scenarios after implementation.**
+  > **The orchestrator WILL verify evidence files exist before marking task complete.**
 
   \\\`\\\`\\\`
-  Scenario: Create user returns 201 with UUID
-    Tool: Bash (curl)
-    Preconditions: Server running on localhost:8080
+  Scenario: [Happy path \u2014 what SHOULD work]
+    Tool: [Playwright / interactive_bash / Bash (curl)]
+    Preconditions: [Exact setup state]
     Steps:
-      1. curl -s -w "\\n%{http_code}" -X POST http://localhost:8080/api/users \\
-           -H "Content-Type: application/json" \\
-           -d '{"email":"new@test.com","name":"Test User"}'
-      2. Assert: HTTP status is 201
-      3. Assert: response.id matches UUID format
-      4. GET /api/users/{returned-id} \u2192 Assert name equals "Test User"
-    Expected Result: User created and retrievable
-    Evidence: Response bodies captured
-
-  Scenario: Duplicate email returns 409
-    Tool: Bash (curl)
-    Preconditions: User with email "new@test.com" already exists
+      1. [Exact action \u2014 specific command/selector/endpoint, no vagueness]
+      2. [Next action \u2014 with expected intermediate state]
+      3. [Assertion \u2014 exact expected value, not "verify it works"]
+    Expected Result: [Concrete, observable, binary pass/fail]
+    Failure Indicators: [What specifically would mean this failed]
+    Evidence: .sisyphus/evidence/task-{N}-{scenario-slug}.{ext}
+
+  Scenario: [Failure/edge case \u2014 what SHOULD fail gracefully]
+    Tool: [same format]
+    Preconditions: [Invalid input / missing dependency / error state]
     Steps:
-      1. Repeat POST with same email
-      2. Assert: HTTP status is 409
-      3. Assert: response.error contains "already exists"
-    Expected Result: Conflict error returned
-    Evidence: Response body captured
+      1. [Trigger the error condition]
+      2. [Assert error is handled correctly]
+    Expected Result: [Graceful failure with correct error message/code]
+    Evidence: .sisyphus/evidence/task-{N}-{scenario-slug}-error.{ext}
   \\\`\\\`\\\`
 
-  **Example \u2014 TUI/CLI (interactive_bash):**
-
-  \\\`\\\`\\\`
-  Scenario: CLI loads config and displays menu
-    Tool: interactive_bash (tmux)
-    Preconditions: Binary built, test config at ./test.yaml
-    Steps:
-      1. tmux new-session: ./my-cli --config test.yaml
-      2. Wait for: "Configuration loaded" in output (timeout: 5s)
-      3. Assert: Menu items visible ("1. Create", "2. List", "3. Exit")
-      4. Send keys: "3" then Enter
-      5. Assert: "Goodbye" in output
-      6. Assert: Process exited with code 0
-    Expected Result: CLI starts, shows menu, exits cleanly
-    Evidence: Terminal output captured
-
-  Scenario: CLI handles missing config gracefully
-    Tool: interactive_bash (tmux)
-    Preconditions: No config file at ./nonexistent.yaml
-    Steps:
-      1. tmux new-session: ./my-cli --config nonexistent.yaml
-      2. Wait for: output (timeout: 3s)
-      3. Assert: stderr contains "Config file not found"
-      4. Assert: Process exited with code 1
-    Expected Result: Meaningful error, non-zero exit
-    Evidence: Error output captured
-  \\\`\\\`\\\`
+  > **Specificity requirements \u2014 every scenario MUST use:**
+  > - **Selectors**: Specific CSS selectors (\`.login-button\`, not "the login button")
+  > - **Data**: Concrete test data (\`"test@example.com"\`, not \`"[email]"\`)
+  > - **Assertions**: Exact values (\`text contains "Welcome back"\`, not "verify it works")
+  > - **Timing**: Wait conditions where relevant (\`timeout: 10s\`)
+  > - **Negative**: At least ONE failure/error scenario per task
+  >
+  > **Anti-patterns (your scenario is INVALID if it looks like this):**
+  > - \u274C "Verify it works correctly" \u2014 HOW? What does "correctly" mean?
+  > - \u274C "Check the API returns data" \u2014 WHAT data? What fields? What values?
+  > - \u274C "Test the component renders" \u2014 WHERE? What selector? What content?
+  > - \u274C Any scenario without an evidence path
 
   **Evidence to Capture:**
-  - [ ] Screenshots in .sisyphus/evidence/ for UI scenarios
-  - [ ] Terminal output for CLI/TUI scenarios
-  - [ ] Response bodies for API scenarios
   - [ ] Each evidence file named: task-{N}-{scenario-slug}.{ext}
+  - [ ] Screenshots for UI, terminal output for CLI, response bodies for API
 
   **Commit**: YES | NO (groups with N)
   - Message: \`type(scope): desc\`
@@ -65631,6 +65632,28 @@ Parallel Speedup: ~40% faster than sequential
 
 ---
 
+## Final Verification Wave (MANDATORY \u2014 after ALL implementation tasks)
+
+> 4 review agents run in PARALLEL. ALL must APPROVE. Rejection \u2192 fix \u2192 re-run.
+
+- [ ] F1. **Plan Compliance Audit** \u2014 \`oracle\`
+  Read the plan end-to-end. For each "Must Have": verify implementation exists (read file, curl endpoint, run command). For each "Must NOT Have": search codebase for forbidden patterns \u2014 reject with file:line if found. Check evidence files exist in .sisyphus/evidence/. Compare deliverables against plan.
+  Output: \`Must Have [N/N] | Must NOT Have [N/N] | Tasks [N/N] | VERDICT: APPROVE/REJECT\`
+
+- [ ] F2. **Code Quality Review** \u2014 \`unspecified-high\`
+  Run \`tsc --noEmit\` + linter + \`bun test\`. Review all changed files for: \`as any\`/\`@ts-ignore\`, empty catches, console.log in prod, commented-out code, unused imports. Check AI slop: excessive comments, over-abstraction, generic names (data/result/item/temp).
+  Output: \`Build [PASS/FAIL] | Lint [PASS/FAIL] | Tests [N pass/N fail] | Files [N clean/N issues] | VERDICT\`
+
+- [ ] F3. **Real Manual QA** \u2014 \`unspecified-high\` (+ \`playwright\` skill if UI)
+  Start from clean state. Execute EVERY QA scenario from EVERY task \u2014 follow exact steps, capture evidence. Test cross-task integration (features working together, not isolation). Test edge cases: empty state, invalid input, rapid actions. Save to \`.sisyphus/evidence/final-qa/\`.
+  Output: \`Scenarios [N/N pass] | Integration [N/N] | Edge Cases [N tested] | VERDICT\`
+
+- [ ] F4. **Scope Fidelity Check** \u2014 \`deep\`
+  For each task: read "What to do", read actual diff (git log/diff). Verify 1:1 \u2014 everything in spec was built (no missing), nothing beyond spec was built (no creep). Check "Must NOT do" compliance. Detect cross-task contamination: Task N touching Task M's files. Flag unaccounted changes.
+  Output: \`Tasks [N/N compliant] | Contamination [CLEAN/N issues] | Unaccounted [CLEAN/N files] | VERDICT\`
+
+---
+
 ## Commit Strategy
 
 | After Task | Message | Files | Verification |
@@ -67540,9 +67563,11 @@ function createChatMessageHandler2(args) {
     }
     const message = output.message;
     if (firstMessageVariantGate.shouldOverride(input.sessionID)) {
-      const variant = input.model && input.agent ? resolveVariantForModel(pluginConfig, input.agent, input.model) : resolveAgentVariant(pluginConfig, input.agent);
-      if (variant !== undefined) {
-        message["variant"] = variant;
+      if (message["variant"] === undefined) {
+        const variant = input.model && input.agent ? resolveVariantForModel(pluginConfig, input.agent, input.model) : resolveAgentVariant(pluginConfig, input.agent);
+        if (variant !== undefined) {
+          message["variant"] = variant;
+        }
       }
       firstMessageVariantGate.markApplied(input.sessionID);
     } else {