From f0506adfeab55c556fc3751f0f6bf868fe5d6b84 Mon Sep 17 00:00:00 2001
From: Danny Kim <temeddix@gmail.com>
Date: Sun, 13 Sep 2026 14:57:01 +0000
Subject: [PATCH] Fail-closed verdict (#3)

Fail-closed verdict (#3)

Co-authored-by: Danny Kim <temeddix@gmail.com>
Co-committed-by: Danny Kim <temeddix@gmail.com>
---
 prompt.md | 16 +++++++++++++---
 run.ts    | 51 +++++++++++++++++++++++++++++----------------------
 2 files changed, 42 insertions(+), 25 deletions(-)

diff --git a/prompt.md b/prompt.md
index b03df2c..ee4cbad 100644
--- a/prompt.md
+++ b/prompt.md
@@ -19,12 +19,22 @@ against the PR's base and head, and make its complete output your final response
 instead of the short comment style above. Check the whole repository against the
 code rules at the end of this prompt, not only the diff; a violation is at least
 Important even when the diff did not cause it. Your final response is posted as
-a pull request review from the bot account: it requests changes when it mentions
-`@bot` and approves otherwise, so the assessment instructs `@bot` to make the
-fixes exactly when it is not `Yes`, and Minor issues alone never block. For UI
+a pull request review from the bot account, and its first line must be exactly
+the template's verdict and nothing else: `Yes`, `No`, or `With fixes`. `Yes`
+approves and the other two request changes; any other first line is posted as a
+plain comment. Minor issues alone never block. When the verdict is not `Yes`,
+the second line is a one-line instruction to `@bot` to make the fixes. For UI
 changes, check that the result is aligned, clean, and pixel-perfect, and that
 included screenshots prove the intended result was achieved.
 
+The review must read at a glance: everything outside `<details>` blocks totals
+under 512 bytes. Only core information stays visible: the verdict, the `@bot`
+line, and the section headings. Anything verbose goes into a `<details>` block
+whose `<summary>` is a few words, such as the `file:line` and title of an issue
+with the what, why, and how inside; the same for each strength, each
+recommendation, the reasoning, and any compliance notes. Details blocks are
+top-level, never inside a list item, because Gitea breaks them there.
+
 For an `issue_comment` or `pull_request_review_comment` event, treat the `body`
 in the triggering comment payload below as the user's exact instruction.
 
diff --git a/run.ts b/run.ts
index 8978dde..0850f86 100644
--- a/run.ts
+++ b/run.ts
@@ -52,19 +52,21 @@ async function postComment(body: string): Promise<void> {
   });
 }
 
-// A pull request event is a review request, so the response becomes a review:
-// changes are requested when the agent asked @bot to fix something, a failed
-// run only comments, and anything else approves.
+// A pull request event is a review request, so the response becomes a review.
+// Its first line is the verdict; anything unexpected only comments, never
+// approves.
+const VERDICTS: Record<string, string> = {
+  Yes: "APPROVED",
+  No: "REQUEST_CHANGES",
+  "With fixes": "REQUEST_CHANGES",
+};
+
 async function postResult(body: string): Promise<void> {
   if (EVENT !== "pull_request") return postComment(body);
-  const event = body.includes("@bot")
-    ? "REQUEST_CHANGES"
-    : body.startsWith("Bot failed:")
-    ? "COMMENT"
-    : "APPROVED";
+  const [verdict] = body.split("\n", 1);
   await gitea(REVIEWER_TOKEN, `repos/${REPO}/pulls/${INDEX}/reviews`, {
     body: stripAnsi(body),
-    event,
+    event: VERDICTS[verdict.trim()] ?? "COMMENT",
   });
 }
 
@@ -122,10 +124,9 @@ async function renderPrompt(): Promise<string> {
 async function runClaude(prompt: string): Promise<string> {
   const token = Deno.env.get("BOT_TOKEN");
   if (!token) {
-    await postComment(
+    throw new Error(
       "Run `claude setup-token` locally and set the `bot-token` action input.",
     );
-    Deno.exit(1);
   }
   await installSuperpowers("claude");
   const claude = new Deno.Command("claude", {
@@ -144,7 +145,7 @@ async function runClaude(prompt: string): Promise<string> {
     clearEnv: true,
     stdout: "piped",
   }).spawn();
-  let result = "Bot failed: no result";
+  let result: { result?: string; subtype: string } | undefined;
   // Print events as they stream so the runner does not kill the job as a zombie.
   const lines = claude.stdout
     .pipeThrough(new TextDecoderStream())
@@ -156,12 +157,13 @@ async function runClaude(prompt: string): Promise<string> {
       const text = part.thinking ?? part.text ?? part.name;
       if (text) console.log(text);
     }
-    if (event.type === "result") {
-      result = event.result ?? `Bot failed: ${event.subtype}`;
-    }
+    if (event.type === "result") result = event;
   }
   await claude.status;
-  return result;
+  if (result?.result === undefined) {
+    throw new Error(`claude ended with ${result?.subtype ?? "no result"}`);
+  }
+  return result.result;
 }
 
 // Posts the device code so a human can finish the login on the persisted home.
@@ -212,9 +214,14 @@ async function runCodex(prompt: string): Promise<string> {
   return await Deno.readTextFile(file);
 }
 
-await configureGitAuthor();
-const prompt = await renderPrompt();
-const result = env("BOT_TYPE") === "claude"
-  ? await runClaude(prompt)
-  : await runCodex(prompt);
-await postResult(result);
+try {
+  await configureGitAuthor();
+  const prompt = await renderPrompt();
+  const result = env("BOT_TYPE") === "claude"
+    ? await runClaude(prompt)
+    : await runCodex(prompt);
+  await postResult(result);
+} catch (error) {
+  await postComment(`Bot failed: ${error}`);
+  throw error;
+}