@tryarcanist/cli 0.1.249 → 0.1.251

package/README.md

@@ -191,7 +191,6 @@ printf "add tests" | arcanist sessions create your-org/your-repo --prompt-stdin
 arcanist sessions create your-org/your-repo - --model gpt-5.5
 arcanist sessions create your-org/your-repo "port to claude" --backend claude_code --model claude-opus-4-8
 arcanist sessions create your-org/your-repo "refactor auth" --reasoning-effort xhigh
-arcanist sessions create your-org/your-repo "verify this automatically" --auto-verify
 arcanist sessions create your-org/your-repo "use the saved login" --browser-identity identity-id
 arcanist sessions create your-org/your-repo "start signed out" --browser-identity none
 arcanist sessions create your-org/your-repo "fix release branch" --base-branch release/2026-06
@@ -203,7 +202,7 @@ arcanist sessions create your-org/your-repo "retry-safe create" --idempotency-ke
 
 `--wait` blocks until the created prompt finishes, prints the resulting PR/branch line in human-readable mode when it is already available, and exits non-zero if the prompt finishes with `status: failed`, making it suitable for cron or other schedulers that alert on command failure. JSON mode waits quietly and prints the create payload after the prompt completes successfully. `--poll-interval <ms>` tunes the completion check frequency.
 
-Use `--auto-verify` to opt the created session into automatic QA verification after PR creation. Auto verification is default-off at session create unless another surface explicitly enables it.
+`--auto-verify` is a deprecated no-op. The verifier child is no longer spawned at publish — QA runs only on an explicit request (GitHub `/qa`, Slack `/qa`, API, or the UI "Verify"/"QA Test PR" button). The flag is accepted for backward compatibility and no longer starts a verifier.
 
 Use `--browser-identity <id|none>` to attach a business browser identity to the session, or pass `none` to explicitly start without one. Access is authorized by the control plane when the session is created.
 
@@ -225,7 +224,7 @@ Sessions always start from a fresh sandbox (the warm sandbox pool was removed).
 `--idempotency-key <uuid>` is for manually retrying a create request that may have reached the server.
 The CLI derives separate session and prompt idempotency keys from the provided value.
 
-`--onboarding` creates an onboarding session: the agent authors the repo's `.arcanist.json` (`appRuntime` runtime contract plus optional `verify` review-preflight), a root `REVIEW.md` with repo-specific review checks (skipped only when the repo yields none), `.arcanist/` runtime files as needed, and `ARCANIST.md` only when the repo has no usable project doc, proves what it can inside its sandbox, and opens the setup PR ready for review. The onboarding behavior is driven by the bridge's canonical onboarding playbook, not the prompt, so `--onboarding` needs no prompt: `arcanist sessions create <repo> --onboarding --wait`. Any prompt supplied alongside `--onboarding` is ignored. Team use; not part of the external API surface.
+`--onboarding` creates an onboarding session: the agent authors the repo's `.arcanist.json` (`appRuntime` runtime contract plus optional legacy `verify` metadata), a root `REVIEW.md` with repo-specific review checks (skipped only when the repo yields none), `.arcanist/` runtime files as needed, and `ARCANIST.md` only when the repo has no usable project doc, proves what it can inside its sandbox, and opens the setup PR ready for review. Zeus discovers and observes review commands itself; the live server-side `repo_review_config.verify_config_json` surface supplies test-coverage lens overrides. The onboarding behavior is driven by the bridge's canonical onboarding playbook, not the prompt, so `--onboarding` needs no prompt: `arcanist sessions create <repo> --onboarding --wait`. Any prompt supplied alongside `--onboarding` is ignored. Team use; not part of the external API surface.
 
 JSON mode returns `{sessionId, sessionUrl?, repoUrl, model?, agentRuntimeBackend?, reasoningEffort?, autoVerify?, baseBranch?, startBranch?, continuePrUrl?, continueMode?, onboarding?, promptId?}`. `sessionUrl` is emitted only when the server returns it; `agentRuntimeBackend` only when `--backend` is passed. With `--wait`, JSON mode also includes best-effort result fields when available: `prUrl?`, `publishedBranch?`, and `lastBranch?`.
 

package/dist/index.js

@@ -1,4 +1,6 @@
 #!/usr/bin/env node
+import { createRequire as __arcanistCreateRequire } from "node:module";
+const require = __arcanistCreateRequire(import.meta.url);
 var __create = Object.create;
 var __defProp = Object.defineProperty;
 var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
@@ -7993,6 +7995,7 @@ function assertSameAttempt(triggeredSessionId, settledSessionId, label) {
 var PHASES = [
   "idle",
   "running",
+  "review_starting",
   "waiting_for_input",
   "finalizing",
   "review_listening",
@@ -8312,6 +8315,9 @@ var MODEL_REGISTRY = [
     // Default codex model after the gpt-5.5 downgrade; carries gpt-5.5's prior
     // "medium" default so default sessions keep the same reasoning effort.
     reasoning: { efforts: ["none", "low", "medium", "high", "xhigh"], default: "medium" },
+    // No cacheWritePerMillion: the pricing page shows a dash in the cache-writes
+    // column for gpt-5.4 and gpt-5.5, so the absence is deliberate, not an
+    // oversight (checked 2026-07-30). The 5.6 family does publish that rate.
     pricing: {
       inputPerMillion: 2.5,
       outputPerMillion: 15,
@@ -8334,17 +8340,21 @@ var MODEL_REGISTRY = [
     contextWindow: 105e4,
     // Verified 2026-07-09 against OpenAI's GPT-5.6 migration guide and model
     // catalog: the alias routes to gpt-5.6-sol and supports max effort. Pricing
-    // mirrors gpt-5.6-sol since the alias resolves to it.
+    // mirrors gpt-5.6-sol since the alias resolves to it. Cache-write rates
+    // verified 2026-07-30 against the Standard tier at
+    // developers.openai.com/api/docs/pricing.
     reasoning: { efforts: ["none", "low", "medium", "high", "xhigh", "max"], default: "medium" },
     pricing: {
       inputPerMillion: 5,
       outputPerMillion: 30,
       cacheReadPerMillion: 0.5,
+      cacheWritePerMillion: 6.25,
       longContext: {
         thresholdTokens: 272e3,
         inputPerMillion: 10,
         outputPerMillion: 45,
-        cacheReadPerMillion: 1
+        cacheReadPerMillion: 1,
+        cacheWritePerMillion: 12.5
       }
     },
     sessionStart: { eligible: true }
@@ -8357,15 +8367,19 @@ var MODEL_REGISTRY = [
     capabilities: { codexToolSearch: true },
     contextWindow: 105e4,
     reasoning: { efforts: ["none", "low", "medium", "high", "xhigh", "max"], default: "medium" },
+    // Cache-write rates verified 2026-07-30 against the Standard tier at
+    // developers.openai.com/api/docs/pricing.
     pricing: {
       inputPerMillion: 5,
       outputPerMillion: 30,
       cacheReadPerMillion: 0.5,
+      cacheWritePerMillion: 6.25,
       longContext: {
         thresholdTokens: 272e3,
         inputPerMillion: 10,
         outputPerMillion: 45,
-        cacheReadPerMillion: 1
+        cacheReadPerMillion: 1,
+        cacheWritePerMillion: 12.5
       }
     },
     sessionStart: { eligible: true }
@@ -8378,15 +8392,20 @@ var MODEL_REGISTRY = [
     capabilities: { codexToolSearch: true },
     contextWindow: 105e4,
     reasoning: { efforts: ["none", "low", "medium", "high", "xhigh", "max"], default: "medium" },
+    // Verified 2026-07-30 against the Standard tier at
+    // developers.openai.com/api/docs/pricing: short and long context input,
+    // cached input, cache writes, and output.
     pricing: {
-      inputPerMillion: 2.5,
-      outputPerMillion: 15,
-      cacheReadPerMillion: 0.25,
+      inputPerMillion: 2,
+      outputPerMillion: 12,
+      cacheReadPerMillion: 0.2,
+      cacheWritePerMillion: 2.5,
       longContext: {
         thresholdTokens: 272e3,
-        inputPerMillion: 5,
-        outputPerMillion: 22.5,
-        cacheReadPerMillion: 0.5
+        inputPerMillion: 4,
+        outputPerMillion: 18,
+        cacheReadPerMillion: 0.4,
+        cacheWritePerMillion: 5
       }
     },
     sessionStart: { eligible: true }
@@ -8395,19 +8414,25 @@ var MODEL_REGISTRY = [
     id: OpenAIModel.GPT56Luna,
     name: "GPT-5.6 Luna",
     provider: "openai",
+    overloadFallback: OpenAIModel.GPT56Terra,
     backends: [CODEX_AGENT_RUNTIME_BACKEND],
     capabilities: { codexToolSearch: true },
     contextWindow: 105e4,
     reasoning: { efforts: ["none", "low", "medium", "high", "xhigh", "max"], default: "medium" },
+    // Verified 2026-07-30 against the Standard tier at
+    // developers.openai.com/api/docs/pricing: short and long context input,
+    // cached input, cache writes, and output.
     pricing: {
-      inputPerMillion: 1,
-      outputPerMillion: 6,
-      cacheReadPerMillion: 0.1,
+      inputPerMillion: 0.2,
+      outputPerMillion: 1.2,
+      cacheReadPerMillion: 0.02,
+      cacheWritePerMillion: 0.25,
       longContext: {
         thresholdTokens: 272e3,
-        inputPerMillion: 2,
-        outputPerMillion: 9,
-        cacheReadPerMillion: 0.2
+        inputPerMillion: 0.4,
+        outputPerMillion: 1.8,
+        cacheReadPerMillion: 0.04,
+        cacheWritePerMillion: 0.5
       }
     },
     sessionStart: { eligible: true }
@@ -9636,7 +9661,8 @@ function projectRetryStatus(data, index) {
     ...typeof data?.scope === "string" ? { scope: data.scope } : {},
     ...typeof data?.maxAttempts === "number" ? { maxAttempts: data.maxAttempts } : {},
     ...typeof data?.reason === "string" ? { reason: data.reason } : {},
-    ...typeof data?.retryAfterMs === "number" ? { retryAfterMs: data.retryAfterMs } : {}
+    ...typeof data?.retryAfterMs === "number" ? { retryAfterMs: data.retryAfterMs } : {},
+    ...typeof data?.fallbackModel === "string" ? { fallbackModel: data.fallbackModel } : {}
   };
 }
 function projectPromptRetrying(data, index) {
@@ -10105,6 +10131,16 @@ function resolveAuthoritativePromptEventsWithDiagnostics(raw) {
   };
 }
 
+// ../../shared/transcript/review-prompt-sections.ts
+var REVIEW_CALIBRATION_SECTION_HEADING = "## Previously declined findings on this repository (calibration)";
+var REVIEW_EXISTING_BOT_FINDINGS_SECTION_HEADING = "## Findings already reported by other reviewers";
+var REVIEW_STACK_NEIGHBOURS_SECTION_HEADING = "## Graphite stack neighbours";
+var REVIEW_PROMPT_ENRICHMENT_SECTION_HEADINGS = [
+  REVIEW_CALIBRATION_SECTION_HEADING,
+  REVIEW_EXISTING_BOT_FINDINGS_SECTION_HEADING,
+  REVIEW_STACK_NEIGHBOURS_SECTION_HEADING
+];
+
 // ../../shared/transcript/prompt-display.ts
 var WRAPPED_CONTENT_BLOCK_RE = /<(user_content|instruction_content)\b[^>]*>[\s\S]*?<\/\1>/gi;
 var SCAFFOLDING_LINE_PATTERNS = [
@@ -10132,6 +10168,25 @@ var SCAFFOLDING_MARKERS = [
   /^Repository:\s+\S+$/im,
   /^\[arcanist:review-loop\b/im
 ];
+var REVIEW_PROMPT_FIRST_LINE_RE = /^Review pull request \S+ at its current head\./;
+function stripReviewPromptEnrichment(text) {
+  const lines = text.split(/\r?\n/);
+  const headingIndexes = [];
+  const cutLineIndexes = [];
+  lines.forEach((line, lineIndex) => {
+    const headingIndex = REVIEW_PROMPT_ENRICHMENT_SECTION_HEADINGS.findIndex((heading) => line.trim() === heading);
+    if (headingIndex === -1) return;
+    headingIndexes.push(headingIndex);
+    cutLineIndexes.push(lineIndex);
+  });
+  if (cutLineIndexes.length === 0) return text.trim();
+  const isCanonicalRun = headingIndexes.every(
+    (headingIndex, position) => position === 0 || headingIndex > (headingIndexes[position - 1] ?? -1)
+  );
+  if (!isCanonicalRun) return text.trim();
+  const cutAt = cutLineIndexes[0] ?? 0;
+  return lines.slice(0, cutAt).join("\n").trim();
+}
 function promptContainsScaffolding(text) {
   return SCAFFOLDING_MARKERS.some((pattern) => pattern.test(text));
 }
@@ -10148,6 +10203,9 @@ function derivePromptDisplayText(prompt) {
   }
   const reply = prompt.replyToText?.trim();
   if (reply && !promptContainsScaffolding(reply)) return reply;
+  if (REVIEW_PROMPT_FIRST_LINE_RE.test(prompt.prompt.trimStart())) {
+    return stripReviewPromptEnrichment(prompt.prompt.trim());
+  }
   if (promptContainsScaffolding(prompt.prompt)) {
     const recovered = stripPromptScaffolding(prompt.prompt);
     if (recovered) return recovered;
@@ -10166,6 +10224,7 @@ var ERROR_CODES = [
   "aborted",
   "rate_limit",
   "api_error",
+  "model_overloaded",
   "config_error",
   "failed_edits",
   "memory_enforcement_failed",
@@ -10208,6 +10267,7 @@ var ERROR_CODE_LABELS = {
   aborted: "Stopped by user",
   rate_limit: "Rate limited - please retry shortly",
   api_error: "Model service error",
+  model_overloaded: "Model at capacity",
   config_error: "Configuration error",
   failed_edits: "Edit failure",
   memory_enforcement_failed: "Memory enforcement failed",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tryarcanist/cli",
-  "version": "0.1.249",
+  "version": "0.1.251",
   "description": "CLI for Arcanist — create and manage coding agent sessions",
   "type": "module",
   "bin": {