agent-roast 0.1.0 → 0.1.2

package/README.md

@@ -1,79 +1,68 @@
-# agent-roast 🤖🔥
+<div align="center">
 
-> **Zero-telemetry CLI scorecard auditing your git history for AI coding agent infractions, shortcuts, and panic loops.**
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset="docs/assets/logo-dark.svg">
+  <img src="docs/assets/logo.svg" alt="agent-roast" width="340">
+</picture>
 
-[![npm version](https://img.shields.io/npm/v/agent-roast?color=cb3837&label=npm)](https://www.npmjs.com/package/agent-roast)
-[![telemetry: zero](https://img.shields.io/badge/telemetry-zero%20(offline)-1f9d55)](#privacy--security)
+### Audit your git history for AI coding agent infractions and panic loops
+
+[![npm version](https://img.shields.io/npm/v/agent-roast?logo=npm&logoColor=white&color=cb3837)](https://www.npmjs.com/package/agent-roast)
+[![telemetry: zero (offline)](https://img.shields.io/badge/telemetry-zero%20(offline)-1f9d55)](#privacy)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
 
-Did your AI coding agent (Cursor, Claude Code, Aider) write clean software, or did it silently delete tests and slap `as any` everywhere to pretend everything compiles?
+</div>
 
-Run in any git repository (100% local, zero tokens, zero telemetry):
+`agent-roast` scans recorded git diffs for shortcuts taken by AI coding assistants (Claude Code, Cursor, Aider). It checks for skipped tests, compiler type escapes (`as any`), swallowed errors, and panic revert loops, then outputs a discipline score and line proofs. 100% offline, zero API keys, zero telemetry.
 
 ```bash
-npx agent-roast
+npx agent-roast            # scan AI-attributed commits (last 90 days)
+npx agent-roast --all      # scan entire repository history
+npx agent-roast --verbose  # show commit SHAs and line proofs
 ```
 
----
-
-### What it looks like:
+What a report looks like (a real repository audit, trimmed):
 
 ```text
-╔══════════════════════════════════════════════════════════╗
-║                 🤖 AGENT ROAST DOSSIER                    ║
-╚══════════════════════════════════════════════════════════╝
-
-  Repository:  my-app
-  Commits:     18 AI-attributed / 42 total
-  Code Added:  4,210 lines (AI) / 12,480 lines (total)
-
-  AGENT ARCHETYPE:  The Any Architect
-  Leans on compiler bypasses (`as any`, `@ts-ignore`) rather than strict type modeling.
-
-  DISCIPLINE SCORE: 68 / 100
-  (100 = spotless discipline, 0 = complete shortcut addiction)
-
-  📊 INFRACTIONS BREAKDOWN:
-  ───────────────────────────────────────────────────────
-  ✖  Skipped tests:       4 (it.skip, describe.skip, xit)
-  ▲  Type escapes:        16 (as any, @ts-ignore, eslint-disable)
-  ✖  Swallowed errors:    3 (empty catch {}, except: pass)
-  ↻  Panic fix loops:     2 (quick fixes < 15m apart)
-
-  ───────────────────────────────────────────────────────
-  📢 Share your score on X: https://twitter.com/intent/tweet?text=...
-  Tip: Run with --verbose to see exact commit SHAs and lines
-```
+agent-roast — git discipline audit (304 AI commits in Soup)
 
----
+SCORE: 55 / 100 · The Panic Looper
+  Prone to rapid-fire fix and revert thrashing cycles when wrestling stubborn bugs.
 
-### What it detects (only added lines, zero false alarms from file renames):
+INFRACTIONS DETECTED (42)
 
-1. **Skipped Tests:** `it.skip`, `describe.skip`, `xit`, `test.todo` added when tests fail.
-2. **Type Escapes:** `as any`, `// @ts-ignore`, `// @ts-expect-error`, `# type: ignore`, `eslint-disable`.
-3. **Swallowed Errors:** Empty `catch {}` or `except: pass` blocks hiding production bugs.
-4. **Panic Loops:** Consecutive quickfix/revert commits within 15 minutes by the agent.
+  ● panic-loop (15) — consecutive quick fixes < 15m apart
+      latest: 3df29a1 -> a910f21 (3m apart)
 
----
+  ● type-escape (19) — compiler type bypasses in added lines
+      e.g. src/client.ts:42 — response.body as any
 
-### Options
+  ● test-skip (6) — disabled or skipped test suites
+      e.g. tests/test_outbound.py:18 — pytest.skip(...)
 
-```bash
-npx agent-roast            # Audit AI-attributed commits (Co-Authored-By: Claude, Cursor, Aider, etc.)
-npx agent-roast --all      # Audit ALL commits across the repository
-npx agent-roast --verbose  # Print exact commit SHAs and code lines as proof
-npx agent-roast --since 30d # Audit last 30 days of git history
-npx agent-roast --json     # Output raw JSON metrics for CI pipelines
+  ● swallowed-error (2) — empty catch blocks or except: pass
 ```
 
+<img src="docs/assets/case-study.gif" alt="agent-roast git audit session" width="100%">
+
+---
+
+### What it checks
+
+Only **added lines** (`+`) in source files are scanned — file moves and refactoring never trigger false alarms.
+
+- **Skipped tests:** `it.skip`, `describe.skip`, `xit`, `pytest.skip`, `test.todo` added when tests fail.
+- **Type escapes:** `as any`, `// @ts-ignore`, `// @ts-expect-error`, `# type: ignore`, `eslint-disable`.
+- **Swallowed errors:** Empty `catch {}` or `except: pass` blocks hiding production errors.
+- **Panic fix loops:** Rapid-fire fix and revert commits by the same author within 15 minutes.
+
 ---
 
-### Privacy & Security
+### Privacy
 
-`agent-roast` is strictly **local and deterministic**:
-- Runs streaming `git log` on your machine in milliseconds.
-- **Zero network calls**, zero LLM APIs, zero tracking or analytics.
-- Your source code never leaves your terminal.
+- **100% Local:** Runs streaming `git log` on your machine in <200ms.
+- **Zero Telemetry:** No cloud dependencies, no LLM API calls, no network traffic.
+- **Hermetic:** Source code never leaves your computer.
 
 ---
 

package/dist/{chunk-RFCXDN22.js → chunk-EPPMO4JE.js}

@@ -425,71 +425,81 @@ var MetricsCollector = class {
 // src/renderer.ts
 import pc from "picocolors";
 function generateShareUrl(summary) {
-  const text = `My AI coding agent scored ${summary.score}/100 on @agent_roast! \u{1F916}\u{1F525}
-
-Archetype: "${summary.archetype}"
+  const text = `My AI coding agent scored ${summary.score}/100 on agent-roast.
 
+Archetype: ${summary.archetype}
 Infractions:
 \u2022 ${summary.testSkips.length} skipped tests
 \u2022 ${summary.typeEscapes.length} type escapes
 \u2022 ${summary.panicLoops.length} panic revert loops
 
-Audit your repo with: npx agent-roast`;
+Audit your repo: npx agent-roast`;
   return `https://twitter.com/intent/tweet?text=${encodeURIComponent(text)}`;
 }
 function formatTerminalCard(summary, options) {
   const lines = [];
-  const boxW = 60;
-  const hLine = "\u2550".repeat(boxW - 2);
-  lines.push(pc.cyan(`\u2554${hLine}\u2557`));
-  lines.push(
-    pc.cyan("\u2551") + pc.bold(pc.white("                 \u{1F916} AGENT ROAST DOSSIER                    ")) + pc.cyan("\u2551")
-  );
-  lines.push(pc.cyan(`\u255A${hLine}\u255D`));
   lines.push("");
-  lines.push(`  ${pc.dim("Repository:")}  ${pc.bold(pc.white(options.repoName))}`);
   lines.push(
-    `  ${pc.dim("Commits:")}     ${pc.yellow(summary.aiCommits)} AI-attributed / ${summary.totalCommits} total`
+    `${pc.bold(pc.cyan("agent-roast"))} ${pc.dim("\u2014 git discipline audit")} ${pc.white(options.repoName)}`
   );
   lines.push(
-    `  ${pc.dim("Code Added:")}  ${pc.green(summary.aiLinesAdded)} lines (AI) / ${summary.totalLinesAdded} lines (total)`
+    `  ${pc.dim("commits:")} ${pc.yellow(summary.aiCommits)} AI / ${summary.totalCommits} total   ${pc.dim("lines added:")} ${pc.green(summary.aiLinesAdded)} AI / ${summary.totalLinesAdded} total`
   );
   lines.push("");
   if (!summary.isSufficientData) {
-    lines.push(pc.yellow(`  \u26A0  INSUFFICIENT DATA TO ROAST (< 200 lines added)`));
-    lines.push(pc.dim("     Commit more AI code and run `npx agent-roast` again."));
+    lines.push(pc.yellow(`  insufficient data (< 200 lines added by AI)`));
+    lines.push(pc.dim("  commit more code and run `npx agent-roast` again."));
     lines.push("");
     return lines.join("\n");
   }
   let scoreColor = pc.green;
   if (summary.score < 60) scoreColor = pc.red;
   else if (summary.score < 80) scoreColor = pc.yellow;
-  lines.push(`  ${pc.bold("AGENT ARCHETYPE:")}  ${pc.magenta(pc.bold(summary.archetype))}`);
-  lines.push(`  ${pc.italic(summary.archetypeDescription)}`);
-  lines.push("");
-  lines.push(`  ${pc.bold("DISCIPLINE SCORE:")} ${scoreColor(pc.bold(`${summary.score} / 100`))}`);
   lines.push(
-    pc.dim("  (100 = spotless discipline, 0 = complete shortcut addiction)")
+    `  ${pc.bold("SCORE:")} ${scoreColor(pc.bold(`${summary.score} / 100`))}  ${pc.dim("\xB7")}  ${pc.magenta(pc.bold(summary.archetype))}`
   );
+  lines.push(`  ${pc.dim(summary.archetypeDescription)}`);
   lines.push("");
-  lines.push(pc.bold("  \u{1F4CA} INFRACTIONS BREAKDOWN:"));
-  lines.push(pc.dim("  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500"));
-  const skipIcon = summary.testSkips.length > 0 ? pc.red("\u2716") : pc.green("\u2714");
-  lines.push(
-    `  ${skipIcon}  ${pc.bold("Skipped tests:")}       ${summary.testSkips.length} ${pc.dim("(it.skip, describe.skip, xit)")}`
-  );
-  const escapeIcon = summary.typeEscapes.length > 0 ? pc.yellow("\u25B2") : pc.green("\u2714");
-  lines.push(
-    `  ${escapeIcon}  ${pc.bold("Type escapes:")}        ${summary.typeEscapes.length} ${pc.dim("(as any, @ts-ignore, eslint-disable)")}`
-  );
-  const errorIcon = summary.swallowedErrors.length > 0 ? pc.red("\u2716") : pc.green("\u2714");
-  lines.push(
-    `  ${errorIcon}  ${pc.bold("Swallowed errors:")}    ${summary.swallowedErrors.length} ${pc.dim("(empty catch {}, except: pass)")}`
-  );
-  const panicIcon = summary.panicLoops.length > 0 ? pc.magenta("\u21BB") : pc.green("\u2714");
-  lines.push(
-    `  ${panicIcon}  ${pc.bold("Panic fix loops:")}     ${summary.panicLoops.length} ${pc.dim("(quick fixes < 15m apart)")}`
-  );
+  const totalInfractions = summary.testSkips.length + summary.typeEscapes.length + summary.swallowedErrors.length + summary.panicLoops.length;
+  lines.push(pc.bold(`  INFRACTIONS (${totalInfractions})`));
+  lines.push(pc.dim("  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500"));
+  if (summary.panicLoops.length > 0) {
+    lines.push(
+      `  ${pc.magenta("\u25CF")} ${pc.bold("panic-loop")} (${summary.panicLoops.length})  ${pc.dim("consecutive quick fixes < 15m apart")}`
+    );
+    const sample = summary.panicLoops[0];
+    lines.push(`      ${pc.dim(sample.reason)}`);
+  }
+  if (summary.typeEscapes.length > 0) {
+    lines.push(
+      `  ${pc.yellow("\u25CF")} ${pc.bold("type-escape")} (${summary.typeEscapes.length})  ${pc.dim("compiler bypasses (as any, @ts-ignore)")}`
+    );
+    const sample = summary.typeEscapes[0];
+    if (sample.file) {
+      lines.push(`      ${pc.dim(`${sample.file}: ${sample.lineSnippet || ""}`)}`);
+    }
+  }
+  if (summary.testSkips.length > 0) {
+    lines.push(
+      `  ${pc.red("\u25CF")} ${pc.bold("test-skip")} (${summary.testSkips.length})  ${pc.dim("skipped or disabled tests")}`
+    );
+    const sample = summary.testSkips[0];
+    if (sample.file) {
+      lines.push(`      ${pc.dim(`${sample.file}: ${sample.lineSnippet || ""}`)}`);
+    }
+  }
+  if (summary.swallowedErrors.length > 0) {
+    lines.push(
+      `  ${pc.red("\u25CF")} ${pc.bold("swallowed-error")} (${summary.swallowedErrors.length})  ${pc.dim("empty catch blocks or except: pass")}`
+    );
+    const sample = summary.swallowedErrors[0];
+    if (sample.file) {
+      lines.push(`      ${pc.dim(`${sample.file}: ${sample.lineSnippet || ""}`)}`);
+    }
+  }
+  if (totalInfractions === 0) {
+    lines.push(`  ${pc.green("\u25CF")} ${pc.dim("zero infractions detected across all inspected lines.")}`);
+  }
   lines.push("");
   if (options.verbose) {
     const allInfractions = [
@@ -499,27 +509,22 @@ function formatTerminalCard(summary, options) {
       ...summary.panicLoops
     ];
     if (allInfractions.length > 0) {
-      lines.push(pc.bold("  \u{1F50D} VERBOSE PROOFS (First 10 items):"));
-      lines.push(pc.dim("  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500"));
+      lines.push(pc.bold("  EVIDENCE (First 10 occurrences):"));
+      lines.push(pc.dim("  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500"));
       for (const inf of allInfractions.slice(0, 10)) {
-        const fileLoc = inf.file ? `${inf.file}` : "git commit";
+        const fileLoc = inf.file ? `${inf.file}` : "commit history";
         lines.push(
-          `  \u2022 ${pc.cyan(inf.commitHash.slice(0, 7))} [${inf.type}] ${pc.white(fileLoc)}`
+          `  \u25CF ${pc.cyan(inf.commitHash.slice(0, 7))} [${inf.type}] ${pc.white(fileLoc)}`
         );
         if (inf.lineSnippet) {
-          lines.push(`    ${pc.dim(inf.lineSnippet)}`);
+          lines.push(`      ${pc.dim(`\u2514\u2500 ${inf.lineSnippet}`)}`);
         }
       }
       lines.push("");
     }
   }
-  const shareUrl = generateShareUrl(summary);
-  lines.push(pc.dim("  \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500"));
-  lines.push(
-    `  ${pc.cyan("\u{1F4E2} Share your score on X:")} ${pc.underline(pc.cyan(shareUrl))}`
-  );
   lines.push(
-    `  ${pc.dim("Tip: Run with --verbose to see exact commit SHAs and lines")}`
+    `  ${pc.dim("Share score:")} ${pc.cyan(generateShareUrl(summary))}`
   );
   lines.push("");
   return lines.join("\n");

package/dist/cli.js

@@ -1,7 +1,7 @@
 #!/usr/bin/env node
 import {
   roastRepository
-} from "./chunk-RFCXDN22.js";
+} from "./chunk-EPPMO4JE.js";
 
 // src/cli.ts
 import { Command } from "commander";

package/dist/index.js

@@ -8,7 +8,7 @@ import {
   roastRepository,
   streamGitLog,
   stripStringAndRegexLiterals
-} from "./chunk-RFCXDN22.js";
+} from "./chunk-EPPMO4JE.js";
 export {
   MetricsCollector,
   detectAiAttribution,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "agent-roast",
-  "version": "0.1.0",
+  "version": "0.1.2",
   "description": "Zero-telemetry CLI scorecard analyzing git history for AI coding agent infractions and panic loops",
   "type": "module",
   "main": "./dist/index.js",
@@ -44,7 +44,9 @@
     "picocolors": "^1.1.0"
   },
   "devDependencies": {
+    "@napi-rs/canvas": "^1.0.10",
     "@types/node": "^20.0.0",
+    "gifenc": "^1.0.3",
     "tsup": "^8.0.0",
     "typescript": "^5.0.0",
     "vitest": "^3.0.0"