From b8e626c47ea554c2b164fd01d91eaad9227ac560 Mon Sep 17 00:00:00 2001
From: Jack Felke <jrinnfelke@gmail.com>
Date: Thu, 12 Mar 2026 13:45:02 -0700
Subject: [PATCH] test: add prompt_score unit tests + fix scope scoring bug
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Export scorePrompt for testability
- Fix: long prompts (>100 chars) no longer get max scope score (25→20)
  without explicit scope-bounding keywords
- Add 16 tests covering all scoring dimensions, grading, and feedback
- Test count: 43 → 59
---
 src/tools/prompt-score.ts        |  6 +-
 tests/tools/prompt-score.test.ts | 95 ++++++++++++++++++++++++++++++++
 2 files changed, 99 insertions(+), 2 deletions(-)
 create mode 100644 tests/tools/prompt-score.test.ts

diff --git a/src/tools/prompt-score.ts b/src/tools/prompt-score.ts
index 1cecf01..1468a32 100644
--- a/src/tools/prompt-score.ts
+++ b/src/tools/prompt-score.ts
@@ -40,7 +40,7 @@ interface ScoreResult {
   feedback: string[];
 }
 
-function scorePrompt(text: string): ScoreResult {
+export function scorePrompt(text: string): ScoreResult {
   const feedback: string[] = [];
   let specificity: number;
   let scope: number;
@@ -59,8 +59,10 @@ function scorePrompt(text: string): ScoreResult {
   }
 
   // Scope: bounded task
-  if (/\b(only|just|single|one|specific|this)\b/i.test(text) || text.length > 100) {
+  if (/\b(only|just|single|one|specific|this)\b/i.test(text)) {
     scope = 25;
+  } else if (text.length > 100) {
+    scope = 20;
   } else if (/\b(all|every|entire|whole)\b/i.test(text)) {
     scope = 10;
     feedback.push("🎯 'All/every' is broad — can you narrow the scope?");
diff --git a/tests/tools/prompt-score.test.ts b/tests/tools/prompt-score.test.ts
new file mode 100644
index 0000000..65648f2
--- /dev/null
+++ b/tests/tools/prompt-score.test.ts
@@ -0,0 +1,95 @@
+import { describe, it, expect } from "vitest";
+import { scorePrompt } from "../../src/tools/prompt-score.js";
+
+describe("scorePrompt", () => {
+  it("gives high specificity for file paths", () => {
+    const result = scorePrompt("Fix the bug in src/tools/prompt-score.ts");
+    expect(result.specificity).toBe(25);
+  });
+
+  it("gives high specificity for backtick identifiers", () => {
+    const result = scorePrompt("Rename `scorePrompt` to `evaluatePrompt`");
+    expect(result.specificity).toBe(25);
+  });
+
+  it("gives medium specificity for generic terms", () => {
+    const result = scorePrompt("Fix the component");
+    expect(result.specificity).toBe(15);
+  });
+
+  it("gives low specificity for vague prompts", () => {
+    const result = scorePrompt("Make it work");
+    expect(result.specificity).toBe(5);
+  });
+
+  it("gives high scope for bounded keywords", () => {
+    const result = scorePrompt("Only update the return type");
+    expect(result.actionability).toBeGreaterThan(0);
+    expect(result.scope).toBe(25);
+  });
+
+  it("gives medium scope for long prompts without scope keywords", () => {
+    const longPrompt = "I need you to look at the authentication flow and figure out why the tokens are expiring too quickly and causing users to get logged out repeatedly during their session";
+    const result = scorePrompt(longPrompt);
+    expect(result.scope).toBe(20);
+  });
+
+  it("penalizes unbounded scope keywords", () => {
+    const result = scorePrompt("Fix all the bugs");
+    expect(result.scope).toBe(10);
+  });
+
+  it("gives high actionability for specific verbs", () => {
+    const result = scorePrompt("Extract the validation logic into a helper");
+    expect(result.actionability).toBe(25);
+  });
+
+  it("gives medium actionability for vague verbs", () => {
+    const result = scorePrompt("Make the login page work better");
+    expect(result.actionability).toBe(15);
+  });
+
+  it("gives high done-condition for outcome words", () => {
+    const result = scorePrompt("Fix the test so it should return 200");
+    expect(result.doneCondition).toBe(25);
+  });
+
+  it("gives decent done-condition for questions", () => {
+    const result = scorePrompt("Why is the build failing?");
+    expect(result.doneCondition).toBe(20);
+  });
+
+  it("grades A+ for excellent prompts", () => {
+    const result = scorePrompt(
+      "Rename only the `validateToken` function in src/auth/tokens.ts — it should return a Result type instead of throwing"
+    );
+    expect(result.total).toBeGreaterThanOrEqual(90);
+    expect(result.grade).toBe("A+");
+  });
+
+  it("grades poorly for vague prompts", () => {
+    const result = scorePrompt("do stuff");
+    expect(result.total).toBeLessThan(45);
+    expect(result.grade).toBe("F");
+  });
+
+  it("returns feedback suggestions for weak areas", () => {
+    const result = scorePrompt("do stuff");
+    expect(result.feedback.length).toBeGreaterThan(0);
+    expect(result.feedback.some(f => f.includes("📁"))).toBe(true);
+  });
+
+  it("returns praise for excellent prompts", () => {
+    const result = scorePrompt(
+      "Rename only the `validateToken` function in src/auth/tokens.ts — it should return a Result type"
+    );
+    expect(result.feedback.some(f => f.includes("🏆"))).toBe(true);
+  });
+
+  it("total is sum of all dimensions", () => {
+    const result = scorePrompt("Fix the bug in src/app.ts");
+    expect(result.total).toBe(
+      result.specificity + result.scope + result.actionability + result.doneCondition
+    );
+  });
+});