Layered harness

The mini app organized in layers (controllers, services, and repositories), with instructions for the agent and a hook that stops it from finishing if it breaks the layer rule.

Before you start

The key idea: the code is split into layers, and each layer has one job. The agent has to respect that split, and a hook (a command that the agent’s tool, such as Claude Code or Codex, runs on its own at a specific moment) checks that it does before the agent can finish.

What the architecture is

A layered architecture splits the code into levels. Each level has a clear job, and the levels call each other in one direction only.

Think of a restaurant. The waiter takes the order and brings it to the kitchen. The kitchen decides whether the dish can be made and cooks it. The pantry stores the ingredients. The waiter never goes into the pantry: they ask the kitchen. So if you change how the ingredients are stored, you do not have to change how orders are taken.

In the mini app, Express is the Node library that receives the HTTP requests. The controllers act like the waiter: they receive the request. The services are the kitchen: they apply the library rules. The repositories are the pantry: they keep the data in memory. The presentation, domain, and data layers are the classic names for this split; the Martin Fowler article in the sources covers that topic.

The key rule

Controllers call services, services call repositories, and nothing skips a layer or goes back up. A controller does not read a repository directly, and a repository does not import services, controllers, or app.ts.

Folder What goes here May import
src/controllers/ Express routes: read and validate the request, call a service, and turn error codes into HTTP codes services, models, Express
src/services/ Business rules: book availability, a maximum of three loans, the order of the checks repositories, models
src/repositories/ In-memory storage, one class per entity models
src/models/ Book, member, and loan types, and the AppError error nothing
src/app.ts Composition root: the only file that imports every layer and connects them every layer

The composition root is the place where the pieces are assembled. Here it is src/app.ts, the only file that knows all the layers.

The rules are written in .dependency-cruiser.cjs. dependency-cruiser is a tool that reads the import lines in your code and checks rules about them. There are six:

  • controllers-do-not-import-repositories: a controller does not import a repository.
  • repositories-do-not-import-upper-layers: a repository does not import services, controllers, or app.ts.
  • services-do-not-import-upper-layers: a service does not import controllers or app.ts.
  • models-do-not-import-other-layers: the models do not import any other layer.
  • no-express-outside-controllers: only the controllers and app.ts import Express.
  • no-circular: there are no circular dependencies, that is, two files that import each other.

The tool also counts import type lines, which only bring in types. If it did not count them, a type import would break the rule without anyone noticing.

What the harness includes

The harness has six parts. The security, quality, and clean-code parts are the full versions of the security piece, from day 3 of the series, the quality piece, from day 4, and the clean-code piece, from day 5.

Part Files What it does
Instructions AGENTS.md Explains where each thing goes and which commands to run.
Skill “add an endpoint” .agents/skills/add-endpoint/SKILL.md Steps in order for a new endpoint: repository, service, controller, and a test in a new file.
Architecture .dependency-cruiser.cjs, .agents/hooks/check-architecture-before-stop.mjs The six layer rules, and the stop hook that checks them when the agent finishes.
Security .agents/permissions.json, .agents/hooks/block-dangerous-commands.mjs, .agents/hooks/check-secrets-before-stop.mjs, .secretlintrc.json Permission rules, a hook that blocks dangerous commands before they run, and a secret scanner when the agent finishes.
Quality .agents/hooks/format-after-edit.mjs, .agents/hooks/check-quality-before-stop.mjs, eslint.config.js, prettier.config.mjs A hook that formats each file the agent edits, and a stop hook that runs the formatting check, the linter, the type check, and the tests.
Clean code .agents/hooks/report-duplicates-before-stop.mjs, .agents/skills/clean-code-review/SKILL.md, .jscpd.json A stop hook that reports once each new duplicate jscpd finds, a review skill, and seven rules in AGENTS.md.

The harness is not tied to one agent. The instructions are in AGENTS.md, an open format that several agents read. The hooks, the skill, and the permissions are in the .agents/ folder, instead of the folder of one specific agent. Each agent reads hooks and permissions from its own settings, so those files work once you connect them. The “Connect it to your agent” part below explains how.

A skill is a procedure that the agent loads only when it needs it. The Skills topic explains it. Permission rules stop the agent from using a tool on certain files; Tools and permissions explains them.

The explorer

When it loads, the explorer opens AGENTS.md. Hidden folders such as .agents/ are shown, because the hooks, the permissions, and the skill live there. Dependencies and package-lock.json are not shown.

// Pre-tool hook: connect it to your agent's event that runs before a shell command.
// It reads the hook's JSON input from stdin, finds the shell command and checks it against a short
// list of dangerous patterns. If one matches, it prints the reason to stderr and exits with code 2,
// so the command does not run. Any other command exits with code 0 and no output.
//
// The script is not tied to one agent: it reads no agent-specific environment variable. The command
// is read from `tool_input.command` or from `command`, the two places where agents put it.
//
// This is a safety net for honest mistakes, not a sandbox. A command written in an unusual way, or a
// script that does the same thing from inside a program, gets past it.

const RULES = [
  {
    id: 'recursive-delete',
    reason: 'Recursive deletes cannot be undone.',
    matches: isRecursiveDelete,
  },
  {
    id: 'force-push',
    reason: 'A force push rewrites the history of the remote branch.',
    matches: (words) => isGit(words, 'push') && words.some(isForcePushFlag),
  },
  {
    id: 'hard-reset',
    reason: 'git reset --hard discards uncommitted changes.',
    matches: (words) => isGit(words, 'reset') && words.includes('--hard'),
  },
  {
    id: 'clean-untracked',
    reason: 'git clean -f deletes untracked files, which git cannot restore.',
    matches: (words) => isGit(words, 'clean') && words.some(isForceFlag),
  },
];

// File names that look like .env but are meant to be shared, so they hold no real secrets.
const SHARED_ENV_FILE = /\.(?:example|sample|template)$/u;
const ENV_FILE_REFERENCE = /(?:^|[\s'"=<>(:/\\])(\.env(?:\.[\w.-]+)?)(?=$|[\s'";|&)<>])/gu;

const DOWNLOADERS = new Set([
  'curl',
  'wget',
  'invoke-webrequest',
  'iwr',
  'invoke-restmethod',
  'irm',
]);
const SHELLS = new Set([
  'cmd',
  'sh',
  'bash',
  'zsh',
  'dash',
  'fish',
  'pwsh',
  'powershell',
  'iex',
  'invoke-expression',
]);

readStdin().then((text) => {
  const command = readCommand(text);
  if (command === undefined) {
    return;
  }
  const problem = findProblem(command);
  if (problem !== undefined) {
    process.stderr.write(
      [
        `Blocked by the security hook (${problem.id}): ${problem.reason}`,
        `Command: ${command}`,
        'Do not look for another way to run it. Tell the user what you wanted to do and why.',
        '',
      ].join('\n'),
    );
    process.exitCode = 2;
  }
});

// Reads stdin as a stream. readFileSync(0) would be shorter, but it does not read piped input in
// every Node runtime: in StackBlitz it did not get the input.
function readStdin() {
  return new Promise((resolve, reject) => {
    let text = '';
    process.stdin.setEncoding('utf8');
    process.stdin.on('data', (chunk) => {
      text += chunk;
    });
    process.stdin.on('end', () => resolve(text));
    process.stdin.on('error', reject);
  });
}

// Returns the shell command from the hook input, or undefined when the tool call has none.
// Input that is not JSON blocks the call: a security check that cannot read its input must not
// let the command through.
function readCommand(text) {
  let input;
  try {
    input = JSON.parse(text);
  } catch {
    process.stderr.write('Blocked by the security hook: its input was not valid JSON.\n');
    process.exitCode = 2;
    return undefined;
  }
  const candidate = input?.tool_input?.command ?? input?.command;
  if (Array.isArray(candidate)) {
    return candidate.join(' ');
  }
  return typeof candidate === 'string' && candidate.trim() !== '' ? candidate : undefined;
}

function findProblem(text) {
  const envFile = findEnvFile(text);
  if (envFile !== undefined) {
    return {
      id: 'env-file',
      reason: `${envFile} may hold secrets. The agent does not read or change .env files.`,
    };
  }
  if (pipesDownloadIntoShell(text)) {
    return {
      id: 'download-into-shell',
      reason: 'Piping a download into a shell runs code that nobody has reviewed.',
    };
  }
  for (const segment of splitCommands(text)) {
    const words = toWords(segment);
    const rule = RULES.find((candidate) => candidate.matches(words));
    if (rule !== undefined) {
      return rule;
    }
  }
  return undefined;
}

function findEnvFile(text) {
  for (const match of text.matchAll(ENV_FILE_REFERENCE)) {
    const name = match[1];
    if (!SHARED_ENV_FILE.test(name)) {
      return name;
    }
  }
  return undefined;
}

// Splits on ; && || & and new lines. Pipes stay inside a segment: each side is checked by
// pipesDownloadIntoShell, and the rules look at the words of the whole segment.
function splitCommands(text) {
  return text.split(/;|&&|\|\||&|\r?\n/u).filter((segment) => segment.trim() !== '');
}

// Lowercase words without quotes. A leading sudo, VAR=value assignments and shell wrappers such as
// `bash -lc` are dropped, so `bash -lc "rm -rf dist"` is checked as `rm -rf dist`.
function toWords(segment) {
  const words = segment
    .trim()
    .split(/\s+/u)
    .map((word) => word.replace(/^['"]|['"]$/gu, '').toLowerCase());
  for (;;) {
    if (words[0] === 'sudo' || /^[a-z_][a-z0-9_]*=/u.test(words[0] ?? '')) {
      words.shift();
    } else if (
      SHELLS.has(baseName(words[0])) &&
      /^(?:-[a-z]*c|-command|\/c)$/u.test(words[1] ?? '')
    ) {
      words.splice(0, 2);
    } else {
      return words;
    }
  }
}

function isRecursiveDelete(words) {
  const program = baseName(words[0]);
  const flags = words.slice(1);
  if (program === 'rm') {
    // Short options of rm can be grouped, as in -rf or -fr.
    return flags.some((flag) => flag === '--recursive' || /^-[dfirv]*r[dfirv]*$/u.test(flag));
  }
  if (program === 'remove-item' || program === 'ri') {
    return flags.some((flag) => /^-r(?:e(?:c(?:u(?:r(?:se?)?)?)?)?)?$/u.test(flag));
  }
  if (['rmdir', 'rd', 'del', 'erase'].includes(program)) {
    return flags.includes('/s');
  }
  return false;
}

// True when the words run `git <subcommand>`, also with options before it such as `git -C dir`.
function isGit(words, subcommand) {
  if (baseName(words[0]) !== 'git') {
    return false;
  }
  let index = 1;
  while (words[index]?.startsWith('-')) {
    // -C and -c take a value: skip it too.
    index += words[index] === '-c' ? 2 : 1;
  }
  return words[index] === subcommand;
}

function isForcePushFlag(word) {
  return word.startsWith('--force') || /^-[a-z]*f/u.test(word) || /^\+\S/u.test(word);
}

function isForceFlag(word) {
  return word === '--force' || /^-[a-z]*f/u.test(word);
}

function pipesDownloadIntoShell(text) {
  for (const segment of splitCommands(text)) {
    const programs = segment.split('|').map((part) => baseName(toWords(part)[0]));
    const download = programs.findIndex((program) => DOWNLOADERS.has(program));
    if (download !== -1 && programs.slice(download + 1).some((program) => SHELLS.has(program))) {
      return true;
    }
  }
  return false;
}

// `/usr/bin/rm` and `rm.exe` count as `rm`.
function baseName(word = '') {
  return word
    .split(/[/\\]/u)
    .pop()
    .replace(/\.exe$/u, '');
}
// Stop hook: connect it to your agent's stop event, the moment the agent is about to finish.
// It runs the architecture check (`npm run check:architecture`, dependency-cruiser). If a file
// breaks a layer rule, it prints the report to stderr and exits with code 2. Agents that follow
// this convention read code 2 as "not done yet" and send the report back to the agent. Exit code 0
// with no output means the layers are respected.
//
// The script is not tied to one agent: it reads no agent-specific input or environment variable.
// stdin (the hook's JSON input) is not read: the check does not depend on it. The check runs on
// every stop, so the only way to pass is to fix the code. Whether the agent caps how many times in
// a row it can be sent back depends on the agent that runs the hook.
//
// The tests used to run here too. They moved to the quality piece's stop hook,
// check-quality-before-stop.mjs, so each hook checks one thing.
import { spawnSync } from 'node:child_process';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

// The project root is two folders up from .agents/hooks/. Agents start hooks from different
// folders, so the script does not rely on the current directory.
const scriptDir = path.dirname(fileURLToPath(import.meta.url));
const projectDir = path.resolve(scriptDir, '..', '..');

const COMMAND = 'npm run check:architecture';
const TIMEOUT_MS = 120_000;
const MAX_OUTPUT_CHARS = 4_000;

// shell: true lets Windows run npm.cmd. The command is a fixed string, not user input.
const result = spawnSync(COMMAND, {
  cwd: projectDir,
  shell: true,
  encoding: 'utf8',
  timeout: TIMEOUT_MS,
});

if (result.status !== 0) {
  const output = [result.stdout, result.stderr, result.error?.message]
    .filter(Boolean)
    .join('\n')
    .trim();
  process.stderr.write(
    [
      'Architecture check failed. Move the code to the right layer before you finish.',
      '',
      `$ ${COMMAND}`,
      tail(output),
      '',
    ].join('\n'),
  );
  process.exitCode = 2;
}

// Keeps only the end of long output, where the violations usually are.
function tail(text) {
  if (text.length <= MAX_OUTPUT_CHARS) {
    return text;
  }
  return `...(${text.length - MAX_OUTPUT_CHARS} characters cut)\n${text.slice(-MAX_OUTPUT_CHARS)}`;
}
// Stop hook: connect it to your agent's stop event, the moment the agent is about to finish.
// It runs the formatting check, the linter, the type checker and the tests. If any of them fails,
// it prints the output to stderr and exits with code 2. Agents that follow this convention read
// code 2 as "not done yet" and send the output back to the agent. Exit code 0 with no output means
// every check passed.
//
// The script is not tied to one agent: it reads no agent-specific input or environment variable.
// stdin (the hook's JSON input) is not read: the checks do not depend on it. The checks run on
// every stop, so the only way to pass is to fix the code. Whether the agent caps how many times in
// a row it can be sent back depends on the agent that runs the hook.
import { spawnSync } from 'node:child_process';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

// The project root is two folders up from .agents/hooks/. Agents start hooks from different
// folders, so the script does not rely on the current directory.
const scriptDir = path.dirname(fileURLToPath(import.meta.url));
const projectDir = path.resolve(scriptDir, '..', '..');

// Every command runs even if an earlier one fails, so the agent gets every problem in one report.
// typecheck is here because Vitest runs TypeScript without checking its types: the tests can pass
// while the types are wrong.
const CHECKS = ['npm run format:check', 'npm run lint', 'npm run typecheck', 'npm test'];
const CHECK_TIMEOUT_MS = 120_000;
const MAX_OUTPUT_CHARS = 4_000;

const failures = [];
for (const command of CHECKS) {
  // shell: true lets Windows run npm.cmd. The commands are fixed strings, not user input.
  const result = spawnSync(command, {
    cwd: projectDir,
    shell: true,
    encoding: 'utf8',
    timeout: CHECK_TIMEOUT_MS,
  });
  if (result.status !== 0) {
    const output = [result.stdout, result.stderr, result.error?.message]
      .filter(Boolean)
      .join('\n')
      .trim();
    failures.push({ command, output: tail(output) });
  }
}

if (failures.length > 0) {
  const report = [
    'Quality check failed. Fix these before you finish.',
    'Run `npm run format` to fix formatting. Fix lint, type and test errors in the code, not in the checks.',
    '',
  ];
  for (const failure of failures) {
    report.push(`$ ${failure.command}`, failure.output, '');
  }
  process.stderr.write(report.join('\n'));
  process.exitCode = 2;
}

// Keeps only the end of long output, where the errors and failing assertions usually are.
function tail(text) {
  if (text.length <= MAX_OUTPUT_CHARS) {
    return text;
  }
  return `...(${text.length - MAX_OUTPUT_CHARS} characters cut)\n${text.slice(-MAX_OUTPUT_CHARS)}`;
}
// Stop hook: connect it to your agent's stop event, the moment the agent is about to finish.
// It runs the secret scanner (`npm run check:secrets`, secretlint). If it finds a secret, it prints
// the report to stderr and exits with code 2. Agents that follow this convention read code 2 as
// "not done yet" and send the report back to the agent. Exit code 0 with no output means no secret
// was found.
//
// The script is not tied to one agent: it reads no agent-specific input or environment variable.
// stdin (the hook's JSON input) is not read: the scan does not depend on it.
import { spawnSync } from 'node:child_process';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

// The project root is two folders up from .agents/hooks/. Agents start hooks from different
// folders, so the script does not rely on the current directory.
const scriptDir = path.dirname(fileURLToPath(import.meta.url));
const projectDir = path.resolve(scriptDir, '..', '..');

const COMMAND = 'npm run check:secrets';
const TIMEOUT_MS = 60_000;
const MAX_OUTPUT_CHARS = 4_000;

// shell: true lets Windows run npm.cmd. The command is a fixed string, not user input.
const result = spawnSync(COMMAND, {
  cwd: projectDir,
  shell: true,
  encoding: 'utf8',
  timeout: TIMEOUT_MS,
});

if (result.status !== 0) {
  const output = [result.stdout, result.stderr, result.error?.message]
    .filter(Boolean)
    .join('\n')
    .trim();
  process.stderr.write(
    [
      'Secret check failed. Remove the secret before you finish.',
      'Read it from an environment variable instead, and keep the real value in .env, which is not committed.',
      '',
      `$ ${COMMAND}`,
      tail(output),
      '',
    ].join('\n'),
  );
  process.exitCode = 2;
}

// Keeps only the end of long output, where the findings usually are.
function tail(text) {
  if (text.length <= MAX_OUTPUT_CHARS) {
    return text;
  }
  return `...(${text.length - MAX_OUTPUT_CHARS} characters cut)\n${text.slice(-MAX_OUTPUT_CHARS)}`;
}
// Post-edit hook: connect it to your agent's event that runs after it writes or edits a file.
// It reads the hook's JSON input from stdin, finds the edited files and formats them with Prettier,
// using the project's Prettier config and .prettierignore. If Prettier cannot format a file, for
// example because of a syntax error, it prints the reason to stderr and exits with code 2, so the
// agent sees it. Everything else exits with code 0 and no output: a file outside the project, a
// file that Prettier does not know or ignores, or input without a file path.
//
// The script is not tied to one agent: it reads no agent-specific environment variable. The file
// is read from `tool_input.file_path` or from `file_path`, the two places where agents put it, or
// from the file headers of a patch in `tool_input.command`, for agents that edit with patches.
//
// Formatting here is a convenience, not a gate. Files that the agent changes from the terminal
// never reach this hook, so the stop hook checks the formatting of every file.
import { spawnSync } from 'node:child_process';
import { existsSync } from 'node:fs';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

// The project root is two folders up from .agents/hooks/. Agents start hooks from different
// folders, so the script does not rely on the current directory.
const scriptDir = path.dirname(fileURLToPath(import.meta.url));
const projectDir = path.resolve(scriptDir, '..', '..');

// The hook runs Prettier's own script with node. `npx prettier` would start npm first, which adds
// time to every edit.
const PRETTIER = path.join(projectDir, 'node_modules', 'prettier', 'bin', 'prettier.cjs');
const TIMEOUT_MS = 30_000;

// A patch names each file it adds, changes or moves in a header line such as
// `*** Update File: src/app.ts`.
const PATCH_FILE_HEADER = /^\*\*\* (?:Add File|Update File|Move to): (.+)$/gmu;

readStdin().then((text) => {
  const files = readEditedFiles(text).filter((file) => isInsideProject(file) && existsSync(file));
  if (files.length > 0) {
    format(files);
  }
});

// Reads stdin as a stream. readFileSync(0) would be shorter, but it does not read piped input in
// every Node runtime: in StackBlitz it did not get the input.
function readStdin() {
  return new Promise((resolve, reject) => {
    let text = '';
    process.stdin.setEncoding('utf8');
    process.stdin.on('data', (chunk) => {
      text += chunk;
    });
    process.stdin.on('end', () => resolve(text));
    process.stdin.on('error', reject);
  });
}

// Returns the absolute paths of the edited files, or an empty list when the input has none. Input
// that is not JSON does nothing: a formatter must not get in the agent's way.
function readEditedFiles(text) {
  let input;
  try {
    input = JSON.parse(text);
  } catch {
    return [];
  }
  const base = typeof input?.cwd === 'string' ? input.cwd : projectDir;
  return namedFiles(input)
    .map((name) => name.trim())
    .filter((name) => name !== '')
    .map((name) => path.resolve(base, name));
}

function namedFiles(input) {
  const filePath = input?.tool_input?.file_path ?? input?.file_path;
  if (typeof filePath === 'string') {
    return [filePath];
  }
  const command = input?.tool_input?.command;
  if (typeof command === 'string') {
    return [...command.matchAll(PATCH_FILE_HEADER)].map((match) => match[1]);
  }
  return [];
}

// The hook only touches files inside its own project.
function isInsideProject(file) {
  const relative = path.relative(projectDir, file);
  return relative !== '' && !path.isAbsolute(relative) && relative.split(path.sep)[0] !== '..';
}

function format(files) {
  if (!existsSync(PRETTIER)) {
    report('Prettier is not installed. Run `npm install` in the project.');
    return;
  }
  // --ignore-unknown skips files that Prettier cannot format, such as images.
  const result = spawnSync(
    process.execPath,
    [PRETTIER, '--write', '--ignore-unknown', '--log-level', 'warn', '--no-color', ...files],
    { cwd: projectDir, encoding: 'utf8', timeout: TIMEOUT_MS },
  );
  if (result.status !== 0) {
    const output = [result.stdout, result.stderr, result.error?.message]
      .filter(Boolean)
      .join('\n')
      .trim();
    report(output);
  }
}

function report(output) {
  process.stderr.write(
    ['Formatting failed. The edit was saved, but Prettier could not format it.', output, ''].join(
      '\n',
    ),
  );
  process.exitCode = 2;
}
// Stop hook: connect it to your agent's stop event, the moment the agent is about to finish.
// It runs jscpd on the folders named in the `check:duplicates` script and reports the duplicated
// blocks it has not reported before. Duplicated code is sometimes the right choice, so this is a
// warning, not a gate: each duplicate is reported once. The first time, the hook prints the list to
// stderr and exits with code 2, which agents that follow this convention read as "not done yet". If
// the agent decides to keep a duplicate and tries to finish again, the hook lets it. Exit code 0
// with no output means there is nothing new to report.
//
// The script is not tied to one agent: it reads no agent-specific input or environment variable.
// stdin (the hook's JSON input) is not read: the check does not depend on it.
import { spawnSync } from 'node:child_process';
import { createHash } from 'node:crypto';
import { existsSync, mkdirSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
import path from 'node:path';
import { fileURLToPath } from 'node:url';

// The project root is two folders up from .agents/hooks/. Agents start hooks from different
// folders, so the script does not rely on the current directory.
const scriptDir = path.dirname(fileURLToPath(import.meta.url));
const projectDir = path.resolve(scriptDir, '..', '..');

// node_modules/.cache/ is where Node tools usually keep their caches, and git already ignores it.
// The JSON report and the list of reported duplicates live there, out of the agent's way.
const CACHE_DIR = 'node_modules/.cache/report-duplicates';
const REPORT_FILE = path.join(projectDir, CACHE_DIR, 'jscpd-report.json');
const REPORTED_FILE = path.join(projectDir, CACHE_DIR, 'reported.json');
const JSCPD = path.join(projectDir, 'node_modules', 'jscpd', 'bin', 'jscpd');
const CHECK_TIMEOUT_MS = 60_000;
const MAX_LISTED = 10;

const problem = runJscpd();
if (problem !== null) {
  // A check that cannot run must not pass in silence. The usual cause is a missing install or a
  // folder in check:duplicates that does not exist.
  process.stderr.write(
    `Could not run the duplicate check (\`npm run check:duplicates\`).\n${problem}\n` +
      'Fix the setup, or tell the user if you cannot.\n',
  );
  process.exitCode = 2;
} else {
  const duplicates = readDuplicates();
  const reported = readReported();
  const fresh = duplicates.filter((duplicate) => !reported.has(duplicate.id));
  // Only the duplicates that still exist are remembered. If one is removed and comes back later, it
  // is reported again.
  writeReported(duplicates.map((duplicate) => duplicate.id));
  if (fresh.length > 0) {
    process.stderr.write(warning(fresh));
    process.exitCode = 2;
  }
}

// The folders come from the check:duplicates script in package.json (for example `jscpd src`), so
// the hook and the command a person runs always scan the same code. The hook starts jscpd with node
// instead of npm, which added about 0.7 seconds to every run. Returns null when the report is ready,
// or the reason it is not.
function runJscpd() {
  const [command, ...args] = readCheckDuplicatesScript();
  if (command !== 'jscpd') {
    return 'The check:duplicates script in package.json must run jscpd, for example `jscpd src`.';
  }
  rmSync(REPORT_FILE, { force: true });
  const reportArgs = ['--silent', '--reporters', 'json', '--output', CACHE_DIR];
  const result = spawnSync(process.execPath, [JSCPD, ...args, ...reportArgs], {
    cwd: projectDir,
    encoding: 'utf8',
    timeout: CHECK_TIMEOUT_MS,
  });
  if (result.status === 0 && existsSync(REPORT_FILE)) {
    return null;
  }
  return [result.stdout, result.stderr, result.error?.message].filter(Boolean).join('\n').trim();
}

function readCheckDuplicatesScript() {
  try {
    const packageJson = JSON.parse(readFileSync(path.join(projectDir, 'package.json'), 'utf8'));
    return String(packageJson.scripts?.['check:duplicates'] ?? '')
      .trim()
      .split(/\s+/);
  } catch {
    return [];
  }
}

// Each duplicate gets an id from its two files and its code, not its line numbers, so an edit
// above it does not make it new.
function readDuplicates() {
  const report = JSON.parse(readFileSync(REPORT_FILE, 'utf8'));
  return report.duplicates.map((duplicate) => {
    const first = location(duplicate.firstFile);
    const second = location(duplicate.secondFile);
    const files = [first.file, second.file].sort().join('\n');
    const id = createHash('sha256').update(`${files}\n${duplicate.fragment}`).digest('hex');
    return { id, first, second, lines: duplicate.lines };
  });
}

// jscpd writes paths with the separator of the operating system. The agent gets forward slashes.
function location(file) {
  return { file: file.name.replaceAll('\\', '/'), start: file.start, end: file.end };
}

function readReported() {
  try {
    return new Set(JSON.parse(readFileSync(REPORTED_FILE, 'utf8')));
  } catch {
    return new Set();
  }
}

function writeReported(ids) {
  mkdirSync(path.dirname(REPORTED_FILE), { recursive: true });
  writeFileSync(REPORTED_FILE, `${JSON.stringify(ids, null, 2)}\n`);
}

function warning(fresh) {
  const lines = [
    'Duplicated code: a warning, not a failure. jscpd found the same code in more than one place:',
    '',
  ];
  for (const { first, second, lines: count } of fresh.slice(0, MAX_LISTED)) {
    lines.push(
      `- ${first.file}:${first.start}-${first.end} and ${second.file}:${second.start}-${second.end} (${count} lines)`,
    );
  }
  if (fresh.length > MAX_LISTED) {
    lines.push(`- and ${fresh.length - MAX_LISTED} more. Run \`npm run check:duplicates\`.`);
  }
  lines.push(
    '',
    'Look at each one before you finish:',
    '- If both copies hold the same rule and would change together, move it into one function and call it from both places.',
    '- If they only look alike, or a shared function would need a flag or extra parameters to tell its callers apart, keep the copies.',
    'Tell the user which duplicates you kept and why. Each duplicate is reported once: when you try to finish again, this hook lets you.',
    '',
  );
  return lines.join('\n');
}
---
name: add-endpoint
description: Add a new HTTP endpoint to the library API in layer order (repository, service, controller, then a new test file). Use when asked to add a route, an endpoint, or an API operation.
---

# Add an endpoint

Follow these steps in this order. Do not start with the controller.

1. **Repository** (`src/repositories/`). Add the storage method the endpoint needs, such as `findAll()` or `setReturnedAt()`. A repository imports only `src/models/`. It knows nothing about HTTP or about business rules.
2. **Service** (`src/services/`). Add the business rule as a method. Check existence and the rules here, in the order the spec requires. Throw `AppError` with a code from `src/models/app-error.ts`. Do not set HTTP status codes in a service.
3. **Controller** (`src/controllers/`). Add the route to the router of its entity. Read the body with `readInput` and `readText` from `http.ts`, throw `AppError('INVALID_INPUT', ...)` when the input is wrong, call the service, and wrap the handler in `try`/`catch` that calls `handleError`. If the router is new, mount it in `createApp()` in `src/app.ts`. If you add an error code, give it a status in `STATUS_BY_CODE` in `src/controllers/http.ts`.
4. **Test** (`test/`). Create a new file, for example `test/<feature>.test.ts`. Test the endpoint over HTTP with Supertest against `createApp()`. Test a service directly when its rule is complex. Never edit `test/library.test.ts`.
5. **Verify.** Run `npm run verify` in this folder. If `check:architecture` fails, the import is in the wrong place: move the code to the correct layer. Do not change the rules in `.dependency-cruiser.cjs` to make the check pass.
---
name: clean-code-review
description: Review your own changes for duplicated code, names and responsibilities before you say a task is done. Use at the end of any task that added or changed code, or when the user asks for a clean-code review.
---

# Clean-code review

Review only the code this task changed. List it with `git status` and `git diff`. Change only what is clearly better: a review that adds layers, helpers or defensive code that nobody asked for makes the code worse.

1. **Duplicates.** Run `npm run check:duplicates`. jscpd only finds copies with the same tokens, so it misses the same logic written with other names. For each function you added, search the code for its main domain word (for example, `grep -rn "loan" src`) and read what is already there. For each duplicate, extract a shared function only if both copies hold the same rule and would change together. If they only look alike, or a shared function would need a flag or extra parameters to tell its callers apart, keep the copies.
2. **Names.** Each new file, function and variable says what it is or does, with the words of the domain (book, member, loan). Rename vague names such as `data`, `info`, `item`, `handle`, `process`, `utils` or `temp`. A boolean reads as a question: `isAvailable`, `hasActiveLoans`.
3. **Responsibilities.** Each new function does one thing. If you need "and" to describe it, split it. Keep each part in the folder where the project's `AGENTS.md` says it goes.
4. **Scope.** Remove what the task did not ask for: unused parameters and options, abstractions written for a future need, commented-out code, and changes to files the task did not need.
5. **Verify.** If you changed anything, run `npm run verify` again.
6. **Report.** Tell the user in a few lines what this review changed, and which duplicates you kept and why.
{
  "description": "Permission rules for the agent. No agent reads this file on its own: copy the rules into the permission settings of the agent you use. deny: the agent never reads these files. ask: the agent asks you before running these commands.",
  "deny": {
    "read": [".env", ".env.*"]
  },
  "ask": {
    "commands": ["git push", "npm install", "npm i", "npm add"]
  }
}
import { Router } from 'express';
import { AppError } from '../models/app-error';
import type { BookService } from '../services/book-service';
import { handleError, readInput, readText } from './http';

export function createBooksRouter(books: BookService): Router {
  const router = Router();

  router.get('/books', (_req, res) => {
    res.json(books.list());
  });

  router.post('/books', (req, res, next) => {
    try {
      const input = readInput(req);
      const title = readText(input.title);
      const author = readText(input.author);
      if (title === null || author === null) {
        throw new AppError('INVALID_INPUT', 'Title and author are required.');
      }
      res.status(201).json(books.add(title, author));
    } catch (error) {
      handleError(error, res, next);
    }
  });

  router.get('/books/:id', (req, res, next) => {
    try {
      res.json(books.get(req.params.id));
    } catch (error) {
      handleError(error, res, next);
    }
  });

  return router;
}
import type { NextFunction, Request, Response } from 'express';
import { AppError, type ErrorCode } from '../models/app-error';

// The only place where error codes become HTTP statuses. Services never see these numbers.
const STATUS_BY_CODE: Record<ErrorCode, number> = {
  INVALID_INPUT: 400,
  BOOK_NOT_FOUND: 404,
  MEMBER_NOT_FOUND: 404,
  LOAN_NOT_FOUND: 404,
  BOOK_ALREADY_LOANED: 409,
  LOAN_LIMIT_REACHED: 409,
  LOAN_ALREADY_RETURNED: 409,
};

export function sendError(res: Response, status: number, code: string, message: string): void {
  res.status(status).json({ error: { code, message } });
}

// Answers an AppError with its mapped status. Any other error goes to the error middleware.
export function handleError(error: unknown, res: Response, next: NextFunction): void {
  if (error instanceof AppError) {
    sendError(res, STATUS_BY_CODE[error.code], error.code, error.message);
    return;
  }
  next(error);
}

export function readInput(req: Request): Record<string, unknown> {
  const body: unknown = req.body;
  return isRecord(body) ? body : {};
}

// Returns the trimmed text, or null when the value is not a string or is blank.
export function readText(value: unknown): string | null {
  if (typeof value !== 'string') {
    return null;
  }
  const text = value.trim();
  return text === '' ? null : text;
}

function isRecord(value: unknown): value is Record<string, unknown> {
  return typeof value === 'object' && value !== null && !Array.isArray(value);
}
import { Router } from 'express';
import { AppError } from '../models/app-error';
import type { LoanService } from '../services/loan-service';
import { handleError, readInput, readText } from './http';

export function createLoansRouter(loans: LoanService): Router {
  const router = Router();

  router.get('/loans', (_req, res) => {
    res.json(loans.list());
  });

  router.post('/loans', (req, res, next) => {
    try {
      const input = readInput(req);
      const bookId = readText(input.bookId);
      const memberId = readText(input.memberId);
      if (bookId === null || memberId === null) {
        throw new AppError('INVALID_INPUT', 'bookId and memberId are required.');
      }
      res.status(201).json(loans.lend(bookId, memberId));
    } catch (error) {
      handleError(error, res, next);
    }
  });

  router.post('/loans/:id/return', (req, res, next) => {
    try {
      res.json(loans.returnLoan(req.params.id));
    } catch (error) {
      handleError(error, res, next);
    }
  });

  return router;
}
import { Router } from 'express';
import { AppError } from '../models/app-error';
import type { MemberService } from '../services/member-service';
import { handleError, readInput, readText } from './http';

export function createMembersRouter(members: MemberService): Router {
  const router = Router();

  router.get('/members', (_req, res) => {
    res.json(members.list());
  });

  router.post('/members', (req, res, next) => {
    try {
      const name = readText(readInput(req).name);
      if (name === null) {
        throw new AppError('INVALID_INPUT', 'Name is required.');
      }
      res.status(201).json(members.add(name));
    } catch (error) {
      handleError(error, res, next);
    }
  });

  return router;
}
// Errors raised by services and by controllers for bad input. They carry a code, never an HTTP
// status: the controllers map each code to a status (see src/controllers/http.ts).

export type ErrorCode =
  | 'INVALID_INPUT'
  | 'BOOK_NOT_FOUND'
  | 'MEMBER_NOT_FOUND'
  | 'LOAN_NOT_FOUND'
  | 'BOOK_ALREADY_LOANED'
  | 'LOAN_LIMIT_REACHED'
  | 'LOAN_ALREADY_RETURNED';

export class AppError extends Error {
  readonly code: ErrorCode;

  constructor(code: ErrorCode, message: string) {
    super(message);
    this.name = 'AppError';
    this.code = code;
  }
}
// Entity shapes used by every layer. This file imports nothing.

export interface Book {
  id: string;
  title: string;
  author: string;
  available: boolean; // false while the book has an active loan
}

export interface Member {
  id: string;
  name: string;
}

export interface Loan {
  id: string;
  bookId: string;
  memberId: string;
  loanedAt: string; // ISO 8601
  returnedAt: string | null; // null while the loan is active
}
import type { Book } from '../models/entities';

// In-memory storage for books. Reads return copies, so a caller cannot change stored data by
// accident. Only the methods here change the stored books.
export class BookRepository {
  private readonly books: Book[] = [];
  private lastId = 0;

  create(fields: { title: string; author: string }): Book {
    const book: Book = { id: String(++this.lastId), ...fields, available: true };
    this.books.push(book);
    return { ...book };
  }

  findById(id: string): Book | undefined {
    const book = this.books.find((item) => item.id === id);
    return book && { ...book };
  }

  findAll(): Book[] {
    return this.books.map((book) => ({ ...book }));
  }

  // Does nothing for an unknown id. Services check that the book exists before calling this.
  setAvailable(id: string, available: boolean): void {
    const book = this.books.find((item) => item.id === id);
    if (book) {
      book.available = available;
    }
  }
}
import type { Loan } from '../models/entities';

// In-memory storage for loans. Reads return copies, like the other repositories.
export class LoanRepository {
  private readonly loans: Loan[] = [];
  private lastId = 0;

  create(fields: { bookId: string; memberId: string; loanedAt: string }): Loan {
    const loan: Loan = { id: String(++this.lastId), ...fields, returnedAt: null };
    this.loans.push(loan);
    return { ...loan };
  }

  findById(id: string): Loan | undefined {
    const loan = this.loans.find((item) => item.id === id);
    return loan && { ...loan };
  }

  findAll(): Loan[] {
    return this.loans.map((loan) => ({ ...loan }));
  }

  // Loans that have not been returned yet.
  findActiveByMember(memberId: string): Loan[] {
    return this.loans
      .filter((loan) => loan.memberId === memberId && loan.returnedAt === null)
      .map((loan) => ({ ...loan }));
  }

  // Does nothing for an unknown id. Services check that the loan exists before calling this.
  setReturnedAt(id: string, returnedAt: string): void {
    const loan = this.loans.find((item) => item.id === id);
    if (loan) {
      loan.returnedAt = returnedAt;
    }
  }
}
import type { Member } from '../models/entities';

// In-memory storage for members. Reads return copies, like the other repositories.
export class MemberRepository {
  private readonly members: Member[] = [];
  private lastId = 0;

  create(fields: { name: string }): Member {
    const member: Member = { id: String(++this.lastId), ...fields };
    this.members.push(member);
    return { ...member };
  }

  findById(id: string): Member | undefined {
    const member = this.members.find((item) => item.id === id);
    return member && { ...member };
  }

  findAll(): Member[] {
    return this.members.map((member) => ({ ...member }));
  }
}
import { AppError } from '../models/app-error';
import type { Book } from '../models/entities';
import type { BookRepository } from '../repositories/book-repository';

// Books have no business rules yet. The service still exists so that controllers never
// touch a repository directly.
export class BookService {
  constructor(private readonly books: BookRepository) {}

  add(title: string, author: string): Book {
    return this.books.create({ title, author });
  }

  list(): Book[] {
    return this.books.findAll();
  }

  get(id: string): Book {
    const book = this.books.findById(id);
    if (!book) {
      throw new AppError('BOOK_NOT_FOUND', `Book ${id} does not exist.`);
    }
    return book;
  }
}
import { AppError } from '../models/app-error';
import type { Loan } from '../models/entities';
import type { BookRepository } from '../repositories/book-repository';
import type { LoanRepository } from '../repositories/loan-repository';
import type { MemberRepository } from '../repositories/member-repository';

const MAX_ACTIVE_LOANS = 3;

// The loan rules. The order of the checks is part of the contract: the first failing check wins.
export class LoanService {
  constructor(
    private readonly loans: LoanRepository,
    private readonly books: BookRepository,
    private readonly members: MemberRepository,
  ) {}

  list(): Loan[] {
    return this.loans.findAll();
  }

  lend(bookId: string, memberId: string): Loan {
    const book = this.books.findById(bookId);
    if (!book) {
      throw new AppError('BOOK_NOT_FOUND', `Book ${bookId} does not exist.`);
    }
    const member = this.members.findById(memberId);
    if (!member) {
      throw new AppError('MEMBER_NOT_FOUND', `Member ${memberId} does not exist.`);
    }
    if (!book.available) {
      throw new AppError('BOOK_ALREADY_LOANED', `Book ${bookId} is already loaned.`);
    }
    if (this.loans.findActiveByMember(member.id).length >= MAX_ACTIVE_LOANS) {
      throw new AppError(
        'LOAN_LIMIT_REACHED',
        `Member ${memberId} has ${MAX_ACTIVE_LOANS} active loans.`,
      );
    }

    const loan = this.loans.create({
      bookId: book.id,
      memberId: member.id,
      loanedAt: new Date().toISOString(),
    });
    this.books.setAvailable(book.id, false);
    return loan;
  }

  returnLoan(loanId: string): Loan {
    const loan = this.loans.findById(loanId);
    if (!loan) {
      throw new AppError('LOAN_NOT_FOUND', `Loan ${loanId} does not exist.`);
    }
    if (loan.returnedAt !== null) {
      throw new AppError('LOAN_ALREADY_RETURNED', `Loan ${loan.id} was already returned.`);
    }

    const returnedAt = new Date().toISOString();
    this.loans.setReturnedAt(loan.id, returnedAt);
    this.books.setAvailable(loan.bookId, true);
    return { ...loan, returnedAt };
  }
}
import type { Member } from '../models/entities';
import type { MemberRepository } from '../repositories/member-repository';

// Members have no business rules yet. The service exists for the same reason as BookService.
export class MemberService {
  constructor(private readonly members: MemberRepository) {}

  add(name: string): Member {
    return this.members.create({ name });
  }

  list(): Member[] {
    return this.members.findAll();
  }
}
import express, { type Express, type NextFunction, type Request, type Response } from 'express';
import { createBooksRouter } from './controllers/books-controller';
import { createLoansRouter } from './controllers/loans-controller';
import { createMembersRouter } from './controllers/members-controller';
import { sendError } from './controllers/http';
import { BookRepository } from './repositories/book-repository';
import { LoanRepository } from './repositories/loan-repository';
import { MemberRepository } from './repositories/member-repository';
import { BookService } from './services/book-service';
import { LoanService } from './services/loan-service';
import { MemberService } from './services/member-service';

// Composition root: the only file that imports every layer. It builds new storage on each call,
// so each app starts with empty data and tests do not share state.
export function createApp(): Express {
  const bookRepository = new BookRepository();
  const memberRepository = new MemberRepository();
  const loanRepository = new LoanRepository();

  const app = express();
  app.use(express.json());
  app.use(createBooksRouter(new BookService(bookRepository)));
  app.use(createMembersRouter(new MemberService(memberRepository)));
  app.use(createLoansRouter(new LoanService(loanRepository, bookRepository, memberRepository)));

  app.use((_req, res) => {
    sendError(res, 404, 'NOT_FOUND', 'Route not found.');
  });

  app.use((err: unknown, _req: Request, res: Response, next: NextFunction) => {
    if (res.headersSent) {
      next(err);
      return;
    }
    if (isMalformedJson(err)) {
      sendError(res, 400, 'INVALID_INPUT', 'Request body must be valid JSON.');
      return;
    }
    console.error(err);
    sendError(res, 500, 'INTERNAL_ERROR', 'Unexpected server error.');
  });

  return app;
}

function isMalformedJson(err: unknown): boolean {
  return err instanceof Error && 'type' in err && err.type === 'entity.parse.failed';
}
import { createApp } from './app';

const port = Number(process.env.PORT ?? 3000);

createApp().listen(port, () => {
  console.log(`Library API listening on http://localhost:${port}`);
});
import { spawn } from 'node:child_process';
import { fileURLToPath } from 'node:url';
import { describe, expect, it } from 'vitest';

// These tests run the hook the way an agent does: JSON on stdin, then read the exit code and stderr.

const hook = fileURLToPath(
  new URL('../.agents/hooks/block-dangerous-commands.mjs', import.meta.url),
);

interface HookResult {
  code: number | null;
  stderr: string;
  stdout: string;
}

// spawn writes stdin as a stream, which works in every runtime the template runs in, StackBlitz too.
function runHook(input: string): Promise<HookResult> {
  return new Promise((resolve, reject) => {
    const child = spawn(process.execPath, [hook]);
    let stdout = '';
    let stderr = '';
    child.stdout.setEncoding('utf8').on('data', (chunk: string) => {
      stdout += chunk;
    });
    child.stderr.setEncoding('utf8').on('data', (chunk: string) => {
      stderr += chunk;
    });
    child.on('error', reject);
    child.on('close', (code) => resolve({ code, stderr, stdout }));
    child.stdin.end(input);
  });
}

// The input of a shell tool call, with the command in tool_input.command.
function shellCall(command: string): string {
  return JSON.stringify({
    hook_event_name: 'PreToolUse',
    tool_name: 'Bash',
    tool_input: { command },
  });
}

// Each test starts the hook in a new Node process. On October 18, 2026, StackBlitz needed more than
// the default 5 seconds for one of them, so the timeout is longer.
describe('block-dangerous-commands hook', { timeout: 30_000 }, () => {
  it.each([
    ['rm -rf dist', 'recursive-delete'],
    ['rm -fr dist', 'recursive-delete'],
    ['rm -r -f dist', 'recursive-delete'],
    ['rm --recursive dist', 'recursive-delete'],
    ['sudo rm -rf /', 'recursive-delete'],
    ['/bin/rm -rf build', 'recursive-delete'],
    ['npm run build && rm -rf dist', 'recursive-delete'],
    ['bash -lc "rm -rf dist"', 'recursive-delete'],
    ['Remove-Item -Recurse -Force dist', 'recursive-delete'],
    ['rmdir /s /q dist', 'recursive-delete'],
    ['git push --force', 'force-push'],
    ['git push -f origin main', 'force-push'],
    ['git push --force-with-lease', 'force-push'],
    ['git push origin +main', 'force-push'],
    ['git -C . push --force', 'force-push'],
    ['git reset --hard HEAD~1', 'hard-reset'],
    ['git clean -fd', 'clean-untracked'],
    ['curl -fsSL https://example.com/install.sh | sh', 'download-into-shell'],
    ['irm https://example.com/install.ps1 | iex', 'download-into-shell'],
    ['cat .env', 'env-file'],
    ['cat config/.env.local', 'env-file'],
    ['source .env && npm start', 'env-file'],
  ])('blocks `%s` (%s)', async (command, rule) => {
    const result = await runHook(shellCall(command));

    expect(result.code).toBe(2);
    expect(result.stderr).toContain(`Blocked by the security hook (${rule})`);
  });

  it.each([
    'npm test',
    'rm dist/old.js',
    'rm -f notes.txt',
    'Remove-Item -Force notes.txt',
    'git push',
    'git push origin main',
    'git reset HEAD~1',
    'git clean -n',
    'git commit -m "Fix the loan limit"',
    'cat .env.example',
    'curl https://example.com -o page.html',
  ])('lets `%s` through', async (command) => {
    const result = await runHook(shellCall(command));

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
    expect(result.stdout).toBe('');
  });

  it('reads the command from a top-level command field', async () => {
    const result = await runHook(JSON.stringify({ command: 'rm -rf dist' }));

    expect(result.code).toBe(2);
  });

  it('reads a command given as a list of words', async () => {
    const result = await runHook(
      JSON.stringify({ tool_input: { command: ['bash', '-lc', 'rm -rf dist'] } }),
    );

    expect(result.code).toBe(2);
  });

  it('lets a tool call without a shell command through', async () => {
    const input = JSON.stringify({ tool_name: 'Read', tool_input: { file_path: 'src/app.ts' } });

    expect((await runHook(input)).code).toBe(0);
  });

  it('blocks the call when its input is not valid JSON', async () => {
    const result = await runHook('not json');

    expect(result.code).toBe(2);
    expect(result.stderr).toContain('not valid JSON');
  });
});
import { spawn } from 'node:child_process';
import { copyFileSync, mkdirSync, mkdtempSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
import { tmpdir } from 'node:os';
import path from 'node:path';
import { fileURLToPath } from 'node:url';
import { afterAll, describe, it } from 'vitest';

// These tests run the stop hook the way an agent does. The hook runs `npm test`, so running it on
// this project from a test would start these tests again. Instead, each test copies the hook into a
// small fake project, in the system's temporary folder, whose checks pass or fail on demand. The
// folders are deleted at the end.

const hook = fileURLToPath(
  new URL('../.agents/hooks/check-quality-before-stop.mjs', import.meta.url),
);
const projectDirs: string[] = [];

const CHECK_NAMES = ['format:check', 'lint', 'typecheck', 'test'];

// Each fake check writes its name to ran.txt and fails if its name is in failing.json.
const CHECK_SCRIPT = `import { appendFileSync, readFileSync } from 'node:fs';
const name = process.argv[2];
appendFileSync('ran.txt', name + '\\n');
if (JSON.parse(readFileSync('failing.json', 'utf8')).includes(name)) {
  console.log(name + ' found 1 problem');
  process.exitCode = 1;
}
`;

function createProject(failing: string[]): string {
  const projectDir = mkdtempSync(path.join(tmpdir(), 'quality-hook-'));
  projectDirs.push(projectDir);
  mkdirSync(path.join(projectDir, '.agents', 'hooks'), { recursive: true });
  copyFileSync(hook, path.join(projectDir, '.agents', 'hooks', 'check-quality-before-stop.mjs'));
  const scripts = Object.fromEntries(
    CHECK_NAMES.map((check) => [check, `node check.mjs ${check}`]),
  );
  writeFileSync(path.join(projectDir, 'package.json'), JSON.stringify({ private: true, scripts }));
  writeFileSync(path.join(projectDir, 'check.mjs'), CHECK_SCRIPT);
  writeFileSync(path.join(projectDir, 'failing.json'), JSON.stringify(failing));
  return projectDir;
}

// The hook runs from this project's folder, not the fake project's, as when an agent starts it
// from another folder. spawn is asynchronous, so both tests can run at the same time.
function runHook(projectDir: string): Promise<{ code: number | null; stderr: string }> {
  return new Promise((resolve, reject) => {
    const child = spawn(process.execPath, [
      path.join(projectDir, '.agents', 'hooks', 'check-quality-before-stop.mjs'),
    ]);
    let stderr = '';
    child.stderr.setEncoding('utf8').on('data', (chunk: string) => {
      stderr += chunk;
    });
    child.on('error', reject);
    child.on('close', (code) => resolve({ code, stderr }));
    child.stdin.end('{}');
  });
}

function checksThatRan(projectDir: string): string[] {
  return readFileSync(path.join(projectDir, 'ran.txt'), 'utf8').trim().split('\n');
}

afterAll(() => {
  for (const projectDir of projectDirs) {
    rmSync(projectDir, { recursive: true, force: true });
  }
});

// Each test runs npm four times, which takes a few seconds, so the timeout is longer than usual.
describe.concurrent('check-quality-before-stop hook', { timeout: 60_000 }, () => {
  it('passes when every check passes', async ({ expect }) => {
    const projectDir = createProject([]);

    const result = await runHook(projectDir);

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
    expect(checksThatRan(projectDir)).toEqual(CHECK_NAMES);
  });

  it('fails with code 2 and reports every failing check', async ({ expect }) => {
    const projectDir = createProject(['lint', 'test']);

    const result = await runHook(projectDir);

    expect(result.code).toBe(2);
    expect(result.stderr).toContain('Quality check failed');
    expect(result.stderr).toContain('$ npm run lint');
    expect(result.stderr).toContain('lint found 1 problem');
    expect(result.stderr).toContain('$ npm test');
    expect(result.stderr).toContain('test found 1 problem');
    expect(result.stderr).not.toContain('$ npm run typecheck');
    // A failing check does not stop the ones after it.
    expect(checksThatRan(projectDir)).toEqual(CHECK_NAMES);
  });
});
import { spawnSync } from 'node:child_process';
import { mkdirSync, rmSync, writeFileSync } from 'node:fs';
import { fileURLToPath } from 'node:url';
import { afterEach, describe, expect, it } from 'vitest';

// These tests run the stop hook the way an agent does, against the files of this project.

const hook = fileURLToPath(
  new URL('../.agents/hooks/check-secrets-before-stop.mjs', import.meta.url),
);
// A folder for the fake secret. The test deletes it when it ends.
const leakDir = fileURLToPath(new URL('../tmp-secret-test/', import.meta.url));

function runHook() {
  const result = spawnSync(process.execPath, [hook], { input: '{}', encoding: 'utf8' });
  return { code: result.status, stderr: result.stderr };
}

afterEach(() => {
  rmSync(leakDir, { recursive: true, force: true });
});

// Each test runs npm, which takes a few seconds, so the timeout is longer than the default.
describe('check-secrets-before-stop hook', { timeout: 60_000 }, () => {
  it('passes when no file holds a secret', () => {
    const result = runHook();

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
  });

  it('fails with code 2 when a file holds a token', () => {
    // A fake GitHub token, built here so this test file does not contain one.
    const fakeToken = `ghp_${'FAKE'.repeat(9)}`;
    mkdirSync(leakDir, { recursive: true });
    writeFileSync(`${leakDir}config.ts`, `export const token = '${fakeToken}';\n`);

    const result = runHook();

    expect(result.code).toBe(2);
    expect(result.stderr).toContain('Secret check failed');
    expect(result.stderr).toContain('GITHUB_TOKEN');
    // secretlint masks the secret in its report.
    expect(result.stderr).not.toContain(fakeToken);
  });
});
import { spawn } from 'node:child_process';
import { copyFileSync, mkdirSync, mkdtempSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
import { tmpdir } from 'node:os';
import path from 'node:path';
import { fileURLToPath } from 'node:url';
import { afterEach, beforeEach, describe, expect, it } from 'vitest';

// These tests run the hook the way an agent does: JSON on stdin, then read the exit code, stderr
// and the edited file.
//
// Each test copies the hook into a fake project in the system's temporary folder. Creating and
// deleting files inside this project would race with other checks that read every file at the same
// time, such as a secret scan in a harness.

const hook = fileURLToPath(new URL('../.agents/hooks/format-after-edit.mjs', import.meta.url));
const prettierScript = fileURLToPath(
  new URL('../node_modules/prettier/bin/prettier.cjs', import.meta.url),
);

const UNFORMATTED = 'export const  book = {title:"Dune",\n  year:1965}\n';
const FORMATTED = "export const book = { title: 'Dune', year: 1965 };\n";

let projectDir = '';
let outsideDir = '';

interface HookResult {
  code: number | null;
  stderr: string;
  stdout: string;
}

function createProject(): string {
  const dir = mkdtempSync(path.join(tmpdir(), 'format-hook-'));
  mkdirSync(path.join(dir, '.agents', 'hooks'), { recursive: true });
  copyFileSync(hook, path.join(dir, '.agents', 'hooks', 'format-after-edit.mjs'));
  for (const name of ['prettier.config.mjs', '.prettierignore']) {
    copyFileSync(fileURLToPath(new URL(`../${name}`, import.meta.url)), path.join(dir, name));
  }
  // The hook runs node_modules/prettier/bin/prettier.cjs of its own project. In the fake project,
  // that file loads the Prettier installed in this one.
  const binDir = path.join(dir, 'node_modules', 'prettier', 'bin');
  mkdirSync(binDir, { recursive: true });
  writeFileSync(path.join(binDir, 'prettier.cjs'), `require(${JSON.stringify(prettierScript)});\n`);
  return dir;
}

// spawn writes stdin as a stream, which works in every runtime the template runs in, StackBlitz too.
function runHook(input: string): Promise<HookResult> {
  return new Promise((resolve, reject) => {
    const child = spawn(process.execPath, [
      path.join(projectDir, '.agents', 'hooks', 'format-after-edit.mjs'),
    ]);
    let stdout = '';
    let stderr = '';
    child.stdout.setEncoding('utf8').on('data', (chunk: string) => {
      stdout += chunk;
    });
    child.stderr.setEncoding('utf8').on('data', (chunk: string) => {
      stderr += chunk;
    });
    child.on('error', reject);
    child.on('close', (code) => resolve({ code, stderr, stdout }));
    child.stdin.end(input);
  });
}

function writeFile(dir: string, name: string, content: string): string {
  const file = path.join(dir, name);
  mkdirSync(path.dirname(file), { recursive: true });
  writeFileSync(file, content);
  return file;
}

// The input after an edit tool call, with the file in tool_input.file_path.
function editCall(file: string): string {
  return JSON.stringify({
    hook_event_name: 'PostToolUse',
    tool_name: 'Edit',
    tool_input: { file_path: file },
  });
}

beforeEach(() => {
  projectDir = createProject();
  outsideDir = mkdtempSync(path.join(tmpdir(), 'format-hook-outside-'));
});

afterEach(() => {
  rmSync(projectDir, { recursive: true, force: true });
  rmSync(outsideDir, { recursive: true, force: true });
});

// Each test starts Prettier in a new Node process. On October 18, 2026, StackBlitz needed more than
// the default 5 seconds for some of them, so the timeout is longer.
describe('format-after-edit hook', { timeout: 30_000 }, () => {
  it('formats the file in tool_input.file_path', async () => {
    const file = writeFile(projectDir, 'src/book.ts', UNFORMATTED);

    const result = await runHook(editCall(file));

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
    expect(result.stdout).toBe('');
    expect(readFileSync(file, 'utf8')).toBe(FORMATTED);
  });

  it('formats the file in a top-level file_path', async () => {
    const file = writeFile(projectDir, 'src/book.ts', UNFORMATTED);

    const result = await runHook(JSON.stringify({ file_path: file }));

    expect(result.code).toBe(0);
    expect(readFileSync(file, 'utf8')).toBe(FORMATTED);
  });

  it('reads a relative path from the folder in cwd', async () => {
    const file = writeFile(projectDir, 'src/book.ts', UNFORMATTED);
    const input = { cwd: path.join(projectDir, 'src'), tool_input: { file_path: 'book.ts' } };

    const result = await runHook(JSON.stringify(input));

    expect(result.code).toBe(0);
    expect(readFileSync(file, 'utf8')).toBe(FORMATTED);
  });

  it('formats every file named in the headers of a patch in tool_input.command', async () => {
    const changed = writeFile(projectDir, 'src/book.ts', UNFORMATTED);
    const added = writeFile(projectDir, 'src/member.ts', UNFORMATTED);
    const patch = [
      '*** Begin Patch',
      '*** Update File: src/book.ts',
      '@@',
      '-old line',
      '+new line',
      '*** Add File: src/member.ts',
      '+new file',
      '*** End Patch',
    ].join('\n');
    const input = { cwd: projectDir, tool_name: 'apply_patch', tool_input: { command: patch } };

    const result = await runHook(JSON.stringify(input));

    expect(result.code).toBe(0);
    expect(readFileSync(changed, 'utf8')).toBe(FORMATTED);
    expect(readFileSync(added, 'utf8')).toBe(FORMATTED);
  });

  it('fails with code 2 when Prettier cannot parse the file, and leaves it as it was', async () => {
    const broken = 'export const = ;\n';
    const file = writeFile(projectDir, 'src/broken.ts', broken);

    const result = await runHook(editCall(file));

    expect(result.code).toBe(2);
    expect(result.stderr).toContain('Formatting failed');
    expect(result.stderr).toContain('broken.ts');
    expect(result.stderr).toContain('SyntaxError');
    expect(readFileSync(file, 'utf8')).toBe(broken);
  });

  it('leaves a file outside the project alone', async () => {
    const file = writeFile(outsideDir, 'book.ts', UNFORMATTED);

    const result = await runHook(editCall(file));

    expect(result.code).toBe(0);
    expect(readFileSync(file, 'utf8')).toBe(UNFORMATTED);
  });

  it('leaves a file that Prettier does not know alone', async () => {
    const content = 'some   notes\n';
    const file = writeFile(projectDir, 'notes.unknown', content);

    const result = await runHook(editCall(file));

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
    expect(readFileSync(file, 'utf8')).toBe(content);
  });

  it.each([
    ['a tool call without a file', JSON.stringify({ tool_input: { command: 'npm test' } })],
    ['a file that does not exist', editCall('src/missing.ts')],
    ['input that is not JSON', 'not json'],
  ])('does nothing for %s', async (_case, input) => {
    const result = await runHook(input);

    expect(result.code).toBe(0);
    expect(result.stderr).toBe('');
  });
});
import type { Express } from 'express';
import request from 'supertest';
import { describe, expect, it } from 'vitest';
import { createApp } from '../src/app';

async function createBook(app: Express, title = 'Dune'): Promise<string> {
  const res = await request(app).post('/books').send({ title, author: 'Frank Herbert' });
  return (res.body as { id: string }).id;
}

async function createMember(app: Express, name = 'Ana'): Promise<string> {
  const res = await request(app).post('/members').send({ name });
  return (res.body as { id: string }).id;
}

describe('library API', () => {
  it('creates a book and lists it as available', async () => {
    const app = createApp();

    const created = await request(app)
      .post('/books')
      .send({ title: 'Dune', author: 'Frank Herbert' });
    expect(created.status).toBe(201);
    expect(created.body).toMatchObject({ title: 'Dune', author: 'Frank Herbert', available: true });

    const list = await request(app).get('/books');
    expect(list.status).toBe(200);
    expect(list.body).toContainEqual(
      expect.objectContaining({ id: created.body.id, available: true }),
    );
  });

  it('returns 404 BOOK_NOT_FOUND for an unknown book', async () => {
    const app = createApp();

    const res = await request(app).get('/books/does-not-exist');
    expect(res.status).toBe(404);
    expect(res.body.error.code).toBe('BOOK_NOT_FOUND');
  });

  it('rejects a book without a title with 400 INVALID_INPUT', async () => {
    const app = createApp();

    const res = await request(app).post('/books').send({ author: 'Frank Herbert' });
    expect(res.status).toBe(400);
    expect(res.body.error.code).toBe('INVALID_INPUT');
  });

  it('creates a member and lists it', async () => {
    const app = createApp();

    const created = await request(app).post('/members').send({ name: 'Ana' });
    expect(created.status).toBe(201);
    expect(created.body).toMatchObject({ name: 'Ana' });

    const list = await request(app).get('/members');
    expect(list.status).toBe(200);
    expect(list.body).toContainEqual(expect.objectContaining({ id: created.body.id, name: 'Ana' }));
  });

  it('lends an available book, and the book stops being available', async () => {
    const app = createApp();
    const bookId = await createBook(app);
    const memberId = await createMember(app);

    const loan = await request(app).post('/loans').send({ bookId, memberId });
    expect(loan.status).toBe(201);
    expect(loan.body).toMatchObject({ bookId, memberId, returnedAt: null });

    const book = await request(app).get(`/books/${bookId}`);
    expect(book.status).toBe(200);
    expect(book.body.available).toBe(false);
  });

  it('rejects lending a book that is already loaned with 409 BOOK_ALREADY_LOANED', async () => {
    const app = createApp();
    const bookId = await createBook(app);
    const firstMemberId = await createMember(app, 'Ana');
    const secondMemberId = await createMember(app, 'Luis');

    const first = await request(app).post('/loans').send({ bookId, memberId: firstMemberId });
    expect(first.status).toBe(201);

    const second = await request(app).post('/loans').send({ bookId, memberId: secondMemberId });
    expect(second.status).toBe(409);
    expect(second.body.error.code).toBe('BOOK_ALREADY_LOANED');
  });

  it('rejects a fourth active loan for the same member with 409 LOAN_LIMIT_REACHED', async () => {
    const app = createApp();
    const memberId = await createMember(app);

    for (const title of ['Dune', 'Emma', 'Ulysses']) {
      const bookId = await createBook(app, title);
      const loanRes = await request(app).post('/loans').send({ bookId, memberId });
      expect(loanRes.status).toBe(201);
    }

    const bookId = await createBook(app, 'Beloved');
    const res = await request(app).post('/loans').send({ bookId, memberId });
    expect(res.status).toBe(409);
    expect(res.body.error.code).toBe('LOAN_LIMIT_REACHED');

    const book = await request(app).get(`/books/${bookId}`);
    expect(book.body.available).toBe(true);
  });

  it('returning a loan makes the book available again', async () => {
    const app = createApp();
    const bookId = await createBook(app);
    const memberId = await createMember(app);
    const loanRes = await request(app).post('/loans').send({ bookId, memberId });

    const res = await request(app).post(`/loans/${loanRes.body.id}/return`);
    expect(res.status).toBe(200);
    expect(res.body).toMatchObject({ id: loanRes.body.id, bookId, memberId });
    expect(res.body.returnedAt).toEqual(expect.any(String));

    const book = await request(app).get(`/books/${bookId}`);
    expect(book.body.available).toBe(true);
  });

  it('rejects returning the same loan twice with 409 LOAN_ALREADY_RETURNED', async () => {
    const app = createApp();
    const bookId = await createBook(app);
    const memberId = await createMember(app);
    const loanRes = await request(app).post('/loans').send({ bookId, memberId });
    const loanPath = `/loans/${loanRes.body.id}/return`;

    const first = await request(app).post(loanPath);
    expect(first.status).toBe(200);

    const second = await request(app).post(loanPath);
    expect(second.status).toBe(409);
    expect(second.body.error.code).toBe('LOAN_ALREADY_RETURNED');
  });

  it('rejects a loan for an unknown member with 404 MEMBER_NOT_FOUND', async () => {
    const app = createApp();
    const bookId = await createBook(app);

    const res = await request(app).post('/loans').send({ bookId, memberId: 'does-not-exist' });
    expect(res.status).toBe(404);
    expect(res.body.error.code).toBe('MEMBER_NOT_FOUND');
  });

  it('rejects malformed JSON with 400 INVALID_INPUT', async () => {
    const app = createApp();

    const res = await request(app)
      .post('/books')
      .set('Content-Type', 'application/json')
      .send('{"title": "Dune"');
    expect(res.status).toBe(400);
    expect(res.body.error.code).toBe('INVALID_INPUT');
  });
});
import { describe, expect, it } from 'vitest';
import { AppError } from '../src/models/app-error';
import { BookRepository } from '../src/repositories/book-repository';
import { LoanRepository } from '../src/repositories/loan-repository';
import { MemberRepository } from '../src/repositories/member-repository';
import { LoanService } from '../src/services/loan-service';

// These tests call the service directly, without HTTP. The rules are checked in one place.

function createLoanService() {
  const books = new BookRepository();
  const members = new MemberRepository();
  const loans = new LoanRepository();
  const service = new LoanService(loans, books, members);
  return { books, members, service };
}

// Returns the error code thrown by fn, or undefined when fn does not throw an AppError.
function errorCodeOf(fn: () => unknown): string | undefined {
  try {
    fn();
  } catch (error) {
    return error instanceof AppError ? error.code : undefined;
  }
  return undefined;
}

describe('LoanService', () => {
  it('marks a lent book as unavailable', () => {
    const { books, members, service } = createLoanService();
    const book = books.create({ title: 'Dune', author: 'Frank Herbert' });
    const member = members.create({ name: 'Ana' });

    const loan = service.lend(book.id, member.id);

    expect(loan).toMatchObject({ bookId: book.id, memberId: member.id, returnedAt: null });
    expect(books.findById(book.id)?.available).toBe(false);
  });

  it('throws BOOK_ALREADY_LOANED when the book has an active loan', () => {
    const { books, members, service } = createLoanService();
    const book = books.create({ title: 'Dune', author: 'Frank Herbert' });
    const first = members.create({ name: 'Ana' });
    const second = members.create({ name: 'Luis' });
    service.lend(book.id, first.id);

    expect(errorCodeOf(() => service.lend(book.id, second.id))).toBe('BOOK_ALREADY_LOANED');
  });

  it('refuses a fourth active loan and leaves the book available', () => {
    const { books, members, service } = createLoanService();
    const member = members.create({ name: 'Ana' });
    for (const title of ['Dune', 'Emma', 'Ulysses']) {
      service.lend(books.create({ title, author: 'Author' }).id, member.id);
    }
    const loaned = books.create({ title: 'Beloved', author: 'Toni Morrison' });
    expect(errorCodeOf(() => service.lend(loaned.id, member.id))).toBe('LOAN_LIMIT_REACHED');
    expect(books.findById(loaned.id)?.available).toBe(true);
  });
});
import { spawn } from 'node:child_process';
import { copyFileSync, mkdirSync, mkdtempSync, rmSync, writeFileSync } from 'node:fs';
import { tmpdir } from 'node:os';
import path from 'node:path';
import { fileURLToPath } from 'node:url';
import { afterAll, describe, it } from 'vitest';

// These tests run the stop hook the way an agent does, then read its exit code and stderr. Each
// test copies the hook and .jscpd.json into a small fake project in the system's temporary folder
// and writes some source files. The hook runs the real jscpd installed in this project. The folders
// are deleted at the end.

const hook = fileURLToPath(
  new URL('../.agents/hooks/report-duplicates-before-stop.mjs', import.meta.url),
);
const jscpdConfig = fileURLToPath(new URL('../.jscpd.json', import.meta.url));
const jscpdBin = fileURLToPath(new URL('../node_modules/jscpd/bin/jscpd', import.meta.url));
const projectDirs: string[] = [];

// A function that validates a text field, and the same function copied under another name. jscpd
// sees the copy: only the function name changed.
const READ_TITLE = `export function readTitle(input: unknown): string {
  if (typeof input !== 'string') {
    throw new Error('The value must be text');
  }
  const text = input.trim();
  if (text.length === 0 || text.length > 200) {
    throw new Error('The value must have between 1 and 200 characters');
  }
  return text;
}
`;
const READ_NAME_COPY = READ_TITLE.replace('readTitle', 'readName');
// The same logic with every name and message changed. jscpd compares tokens, so it does not see it.
const READ_NAME_RENAMED = `export function readName(value: unknown): string {
  if (typeof value !== 'string') {
    throw new Error('A name is text');
  }
  const name = value.trim();
  if (name.length === 0 || name.length > 200) {
    throw new Error('A name has between 1 and 200 characters');
  }
  return name;
}
`;
const COUNT_LOANS = `export function countActiveLoans(loans: { returnedAt: string | null }[]): number {
  return loans.filter((loan) => loan.returnedAt === null).length;
}
`;

function createProject(checkDuplicates = 'jscpd src'): string {
  const projectDir = mkdtempSync(path.join(tmpdir(), 'duplicates-hook-'));
  projectDirs.push(projectDir);
  mkdirSync(path.join(projectDir, '.agents', 'hooks'), { recursive: true });
  copyFileSync(
    hook,
    path.join(projectDir, '.agents', 'hooks', 'report-duplicates-before-stop.mjs'),
  );
  copyFileSync(jscpdConfig, path.join(projectDir, '.jscpd.json'));
  const scripts = { 'check:duplicates': checkDuplicates };
  writeFileSync(path.join(projectDir, 'package.json'), JSON.stringify({ private: true, scripts }));
  // The hook runs node_modules/jscpd/bin/jscpd of its own project. In the fake project, that file
  // loads the jscpd installed in this one.
  const binDir = path.join(projectDir, 'node_modules', 'jscpd', 'bin');
  mkdirSync(binDir, { recursive: true });
  writeFileSync(path.join(binDir, 'jscpd'), `require(${JSON.stringify(jscpdBin)});\n`);
  mkdirSync(path.join(projectDir, 'src'));
  return projectDir;
}

function writeSource(projectDir: string, name: string, content: string): void {
  writeFileSync(path.join(projectDir, 'src', name), content);
}

// The hook runs from this project's folder, not the fake project's, as when an agent starts it
// from another folder. spawn is asynchronous, so the tests can run at the same time.
function runHook(projectDir: string): Promise<{ code: number | null; stderr: string }> {
  return new Promise((resolve, reject) => {
    const child = spawn(process.execPath, [
      path.join(projectDir, '.agents', 'hooks', 'report-duplicates-before-stop.mjs'),
    ]);
    let stderr = '';
    child.stderr.setEncoding('utf8').on('data', (chunk: string) => {
      stderr += chunk;
    });
    child.on('error', reject);
    child.on('close', (code) => resolve({ code, stderr }));
    child.stdin.end('{}');
  });
}

afterAll(() => {
  for (const projectDir of projectDirs) {
    rmSync(projectDir, { recursive: true, force: true });
  }
});

// Each run starts jscpd, which takes about a second on a computer and much longer in StackBlitz,
// so the timeout is longer than usual. The hook starts jscpd without npm, which keeps each run light
// enough to run the tests at the same time there too.
describe.concurrent('report-duplicates-before-stop hook', { timeout: 120_000 }, () => {
  it('passes in silence when there is no duplicated code', async ({ expect }) => {
    const projectDir = createProject();
    writeSource(projectDir, 'books.ts', READ_TITLE);
    writeSource(projectDir, 'loans.ts', COUNT_LOANS);

    const result = await runHook(projectDir);

    expect(result.stderr).toBe('');
    expect(result.code).toBe(0);
  });

  it('reports a copied function once, with code 2 and both locations', async ({ expect }) => {
    const projectDir = createProject();
    writeSource(projectDir, 'books.ts', READ_TITLE);
    writeSource(projectDir, 'members.ts', READ_NAME_COPY);

    const first = await runHook(projectDir);
    // The agent keeps the copy and tries to finish again: the hook lets it.
    const second = await runHook(projectDir);

    expect(first.code).toBe(2);
    expect(first.stderr).toContain('Duplicated code: a warning, not a failure.');
    expect(first.stderr).toMatch(/src\/books\.ts:\d+-\d+/);
    expect(first.stderr).toMatch(/src\/members\.ts:\d+-\d+/);
    expect(first.stderr).toContain('Each duplicate is reported once');
    expect(second.stderr).toBe('');
    expect(second.code).toBe(0);
  });

  it('reports a new duplicate, and a duplicate that was removed and comes back', async ({
    expect,
  }) => {
    const projectDir = createProject();
    writeSource(projectDir, 'books.ts', READ_TITLE);
    writeSource(projectDir, 'members.ts', READ_NAME_COPY);
    await runHook(projectDir);

    // The copy in members.ts goes away and a new one appears in authors.ts.
    writeSource(projectDir, 'members.ts', COUNT_LOANS);
    writeSource(projectDir, 'authors.ts', READ_TITLE.replace('readTitle', 'readAuthor'));
    const fresh = await runHook(projectDir);
    // The copy in members.ts comes back.
    writeSource(projectDir, 'members.ts', READ_NAME_COPY);
    const back = await runHook(projectDir);

    expect(fresh.code).toBe(2);
    expect(fresh.stderr).toMatch(/src\/authors\.ts:\d+-\d+/);
    expect(fresh.stderr).not.toContain('src/members.ts');
    expect(back.code).toBe(2);
    expect(back.stderr).toMatch(/src\/members\.ts:\d+-\d+/);
  });

  // A limit of jscpd, not of the hook: the clean-code-review skill asks the agent to look for this.
  it('does not see the same logic written with other names', async ({ expect }) => {
    const projectDir = createProject();
    writeSource(projectDir, 'books.ts', READ_TITLE);
    writeSource(projectDir, 'members.ts', READ_NAME_RENAMED);

    const result = await runHook(projectDir);

    expect(result.stderr).toBe('');
    expect(result.code).toBe(0);
  });

  it('fails with code 2 when the duplicate check cannot run', async ({ expect }) => {
    const projectDir = createProject('jscpd missing-folder');
    writeSource(projectDir, 'books.ts', READ_TITLE);

    const result = await runHook(projectDir);

    expect(result.code).toBe(2);
    expect(result.stderr).toContain('Could not run the duplicate check');
    expect(result.stderr).toContain('missing-folder');
  });
});
// Architecture rules for the layers. Each rule is one sentence in plain words, and each rule fails the
// check when it is broken. Run with: npm run check:architecture
//
// Layers, from the top down: controllers -> services -> repositories -> models.
// app.ts and server.ts sit above everything and build the layers together.
/** @type {import('dependency-cruiser').IConfiguration} */
module.exports = {
  forbidden: [
    {
      // Controllers talk to services only. Storage stays behind the service layer.
      name: 'controllers-do-not-import-repositories',
      severity: 'error',
      from: { path: '^src/controllers/' },
      to: { path: '^src/repositories/' },
    },
    {
      // Repositories are the bottom layer: they may import models, and nothing above them.
      name: 'repositories-do-not-import-upper-layers',
      severity: 'error',
      from: { path: '^src/repositories/' },
      to: { path: '^src/(services|controllers)/|^src/(app|server)\\.ts$' },
    },
    {
      // Services hold the business rules. They must not depend on HTTP controllers or on app.ts.
      name: 'services-do-not-import-upper-layers',
      severity: 'error',
      from: { path: '^src/services/' },
      to: { path: '^src/controllers/|^src/(app|server)\\.ts$' },
    },
    {
      // Models are the innermost layer. They import no other layer of this app.
      name: 'models-do-not-import-other-layers',
      severity: 'error',
      from: { path: '^src/models/' },
      to: { path: '^src/(repositories|services|controllers)/|^src/(app|server)\\.ts$' },
    },
    {
      // Express is an HTTP detail. Only controllers and the composition root may use it.
      name: 'no-express-outside-controllers',
      severity: 'error',
      from: { path: '^src/(models|repositories|services)/' },
      to: { path: '^node_modules/(@types/)?express(/|$)' },
    },
    {
      // A cycle means no module can be understood or tested on its own.
      name: 'no-circular',
      severity: 'error',
      from: { pathNot: '^node_modules' },
      to: { circular: true },
    },
  ],
  options: {
    doNotFollow: { path: 'node_modules' },
    tsConfig: { fileName: 'tsconfig.json' },
    // Count "import type" lines too. Without this, a type-only import would hide a broken rule.
    tsPreCompilationDeps: true,
  },
};
node_modules/
.env
.env.*
!.env.example
{
  "format": ["typescript", "javascript"],
  "minTokens": 50,
  "minLines": 5,
  "noTips": true
}
node_modules/
package-lock.json
{
  "rules": [{ "id": "@secretlint/secretlint-rule-preset-recommend" }]
}
# AGENTS.md

## Project

A small library REST API with books, members and loans. TypeScript, Node 22 or later, Express 5. Data lives in memory. The full contract (entities, endpoints, error codes, business rules) is in `../MINI_APP_SPEC.md`.

This folder is the **layered harness**: the same app as `starting-point`, organized in layers. Requests flow down through the layers. Nothing skips a layer, and nothing goes back up.

Use npm in this folder, not pnpm.

## Layers

| Folder              | What goes here                                                                     | May import                      | Never imports                                  |
| ------------------- | ---------------------------------------------------------------------------------- | ------------------------------- | ---------------------------------------------- |
| `src/models/`       | Entity types (`Book`, `Member`, `Loan`) and `AppError`                             | nothing                         | any other layer, `express`                     |
| `src/repositories/` | In-memory storage, one class per entity                                            | `models`                        | `services`, `controllers`, `app.ts`, `express` |
| `src/services/`     | Business rules: availability, loan limit, the check order                          | `repositories`, `models`        | `controllers`, `app.ts`, `express`             |
| `src/controllers/`  | Express routers: read and validate the request, call a service, map errors to HTTP | `services`, `models`, `express` | `repositories`, `app.ts`                       |
| `src/app.ts`        | Composition root: builds the layers, JSON parsing, 404 and error middleware        | every layer                     | nothing                                        |
| `src/server.ts`     | Starts `createApp()` on `PORT` (default 3000)                                      | `app.ts`                        | any layer                                      |

## The key rule

Controllers call services. Services call repositories. Never skip a layer (a controller that reads a repository) and never import upward (a repository that imports a service). `npm run check:architecture` checks this with dependency-cruiser, using `.dependency-cruiser.cjs`.

## Where new code goes

- A new entity or error code: `src/models/`. A new error code also needs a status in `STATUS_BY_CODE` in `src/controllers/http.ts`.
- Storage: `src/repositories/`, one class per entity.
- Business rules: `src/services/`. Throw `AppError` with a code. Do not set HTTP status codes here.
- Routes: `src/controllers/`. A new router must be mounted in `createApp()` in `src/app.ts`.
- Tests: a new file in `test/`, for example `test/<feature>.test.ts`.

For a new endpoint, use the skill `add-endpoint` (`.agents/skills/add-endpoint/SKILL.md`). It lists the steps in order: repository, service, controller, test.

## Verification

Run these from this folder:

- `npm test`: the acceptance tests in `test/library.test.ts`, the service tests and the tests of the harness hooks.
- `npm run check:architecture`: the layer rules.
- `npm run check:secrets`: secretlint looks for secrets, such as API keys or tokens, in every file that `.gitignore` does not exclude.
- `npm run typecheck`: TypeScript.
- `npm run lint`: ESLint.
- `npm run format:check`: Prettier. Use `npm run format` to fix formatting.
- `npm run verify`: all of the above, in that order.
- `npm run check:duplicates`: jscpd lists duplicated code in `src/`. It is a report and never fails, so it is not part of `verify`.

Run `npm run verify` before you say a task is done.

When you try to finish, four stop hooks run. `.agents/hooks/check-architecture-before-stop.mjs` runs `npm run check:architecture`, `.agents/hooks/check-secrets-before-stop.mjs` runs `npm run check:secrets`, and `.agents/hooks/check-quality-before-stop.mjs` runs `npm run format:check`, `npm run lint`, `npm run typecheck` and `npm test`. If one of them fails, you cannot finish until it passes. Together they run every check in `npm run verify`. The fourth, `.agents/hooks/report-duplicates-before-stop.mjs`, is a warning described in the `Clean code` section.

## Harness files

The harness keeps its agent files in `.agents/`, not in the folder of one specific agent:

- `.agents/hooks/check-architecture-before-stop.mjs`: the stop hook for the layer rules.
- `.agents/hooks/check-secrets-before-stop.mjs` and `.agents/hooks/block-dangerous-commands.mjs`: the security hooks described in the `Security` section.
- `.agents/hooks/format-after-edit.mjs` and `.agents/hooks/check-quality-before-stop.mjs`: the quality hooks described in the `Quality` section.
- `.agents/hooks/report-duplicates-before-stop.mjs`: the duplicate warning described in the `Clean code` section.
- `.agents/skills/add-endpoint/SKILL.md`: the steps for a new endpoint.
- `.agents/skills/clean-code-review/SKILL.md`: the review before you say a task is done.
- `.agents/permissions.json`: the files you must not read and the commands you must ask about.

Each agent reads hooks and permissions from its own settings, which point to these files. The README explains how.

## Never

- Edit `test/library.test.ts`. It is the contract that every template shares. Do not weaken, skip or delete its tests.
- Change `.dependency-cruiser.cjs` to make a failing check pass. Move the code to the right layer instead.
- Edit the files in `.agents/`, or your agent's hook and permission settings, to get past a failing check. Fix the code instead.

## Security

- Never read, print or change `.env` or `.env.*` files. They hold real secrets. If a task needs a secret, ask the user.
- Never write a real secret (an API key, a token, a password or a private key) in code, tests or docs. Read it from an environment variable and add the variable name, without its value, to `.env.example`.
- Ask the user before you run `git push` or install a dependency (`npm install`, `npm i`, `npm add`).
- Before each shell command, the hook `.agents/hooks/block-dangerous-commands.mjs` blocks recursive deletes, force pushes, `git reset --hard`, `git clean -f`, downloads piped into a shell, and commands that mention a `.env` file. If it blocks a command, do not look for another way to run it. Tell the user what you wanted to do and why.
- When you try to finish, the hook `.agents/hooks/check-secrets-before-stop.mjs` runs `npm run check:secrets`. If it finds a secret, remove it from the file. Do not add the file to `.secretlintignore` or a `secretlint-disable` comment to the code to pass the check.
- Never edit the security hooks, `.secretlintrc.json` or your agent's hook and permission settings to get past a check.

## Quality

- After you write or edit a file, the hook `.agents/hooks/format-after-edit.mjs` formats it with Prettier. You do not need to format by hand. If the hook says it could not format a file, the file usually has a syntax error: fix it.
- When you try to finish, the hook `.agents/hooks/check-quality-before-stop.mjs` runs `npm run format:check`, `npm run lint`, `npm run typecheck` and `npm test`. If one fails, you cannot finish until it passes. Run `npm run format` for formatting problems. Fix lint, type and test errors in the code.
- ESLint limits the size of a function: complexity 10, nesting depth 3, 4 parameters and 50 lines. If you reach a limit, split the function. Do not move the code somewhere else only to pass the check.
- `eslint-disable` comments have no effect here. Do not add them. If a rule is wrong for a case, tell the user. Exceptions go in `eslint.config.js`, where a reviewer sees them, and only the user adds them.
- Never weaken, skip or delete a test to make the checks pass. Never edit the quality hooks, `eslint.config.js`, `prettier.config.mjs`, `.prettierignore` or your agent's hook settings to get past a check.

## Clean code

- Before you write a new function, search the code for one that already does the job, and reuse it.
- Name files, functions and variables with the words of the domain (book, member, loan). Do not use vague names such as `data`, `info`, `handle` or `utils`.
- A function does one thing. If you need "and" to describe what it does, split it.
- Do not merge code only because it looks alike. Extract a shared function when the copies hold the same rule and would change together. If the shared function would need a flag or extra parameters to tell its callers apart, keep the copies.
- Write only what the task needs: no options, parameters or abstractions for later, and no changes to code the task does not touch.
- When you try to finish, the hook `.agents/hooks/report-duplicates-before-stop.mjs` runs jscpd on the folders of the `check:duplicates` script and lists the duplicated blocks it has not reported before. It is a warning: decide on each one with the rule above and tell the user which ones you kept. Never edit `.jscpd.json`, the `check:duplicates` script or the hook to hide a duplicate.
- Before you say a task is done, follow the skill `clean-code-review` (`.agents/skills/clean-code-review/SKILL.md`).

## Conventions

- TypeScript in strict mode. No `any`. Treat request bodies as `unknown` and validate them in controllers.
- IDs are strings created by the repositories. Clients must not assume their format.
- Text fields are trimmed and must not be empty after trimming.
- Keep the error messages short and human-readable. Tests check the `code`, not the message.
import eslint from '@eslint/js';
import globals from 'globals';
import tseslint from 'typescript-eslint';

// A few rules on top of the recommended sets, each with the reason it is here.
export default tseslint.config(
  eslint.configs.recommended,
  tseslint.configs.recommended,
  {
    // Scripts such as the hooks run in Node, so they get Node globals like `process`.
    files: ['**/*.{js,mjs,cjs}'],
    languageOptions: { globals: globals.node },
  },
  {
    linterOptions: {
      // An agent that cannot pass a rule may silence it with an `eslint-disable` comment. With this
      // option ESLint ignores those comments, so an exception has to go in this file, where a
      // reviewer sees it.
      noInlineConfig: true,
    },
    rules: {
      // Type-only imports are erased when the code runs. Marking them makes the dependencies
      // easier to read, and it matches verbatimModuleSyntax in tsconfig.json.
      '@typescript-eslint/consistent-type-imports': 'error',
      // Strict equality avoids silent coercions such as '' == 0.
      eqeqeq: 'error',
      // Agents tend to handle a new case by adding one more branch to the function they are
      // editing. These four limits ask for a new function instead, before it becomes hard to read.
      complexity: ['error', 10],
      'max-depth': ['error', 3],
      'max-params': ['error', 4],
      'max-lines-per-function': ['error', { max: 50, skipBlankLines: true, skipComments: true }],
    },
  },
  {
    // A describe block groups many tests, so its length says nothing about how complex it is.
    files: ['test/**'],
    rules: { 'max-lines-per-function': 'off' },
  },
);
{
  "name": "library-layered",
  "private": true,
  "type": "module",
  "engines": {
    "node": ">=22"
  },
  "scripts": {
    "dev": "tsx watch src/server.ts",
    "start": "tsx src/server.ts",
    "test": "vitest run",
    "typecheck": "tsc --noEmit",
    "check:architecture": "depcruise --config .dependency-cruiser.cjs --output-type err src",
    "check:secrets": "secretlint \"**/*\"",
    "check:duplicates": "jscpd src",
    "lint": "eslint .",
    "format": "prettier --write .",
    "format:check": "prettier --check .",
    "verify": "npm run typecheck && npm run lint && npm run format:check && npm run check:architecture && npm run check:secrets && npm test"
  },
  "dependencies": {
    "express": "5.3.0"
  },
  "devDependencies": {
    "@eslint/js": "10.0.1",
    "@secretlint/secretlint-rule-preset-recommend": "13.0.7",
    "@types/express": "5.0.6",
    "@types/node": "22.20.5",
    "@types/supertest": "7.2.1",
    "dependency-cruiser": "18.6.0",
    "eslint": "10.12.0",
    "globals": "17.13.0",
    "jscpd": "4.3.0",
    "prettier": "3.9.10",
    "secretlint": "13.0.7",
    "supertest": "7.3.1",
    "tsx": "4.23.15",
    "typescript": "5.9.3",
    "typescript-eslint": "8.71.2",
    "vitest": "5.0.4"
  }
}
/** @type {import('prettier').Config} */
export default {
  printWidth: 100,
  semi: true,
  singleQuote: true,
  trailingComma: 'all',
};
# Library layered harness

The library API from [`starting-point`](../starting-point/), organized in layers and protected by a harness. The endpoints, behavior and acceptance tests are the same: `test/library.test.ts` is byte-for-byte identical to the starting point.

This is the first base harness of the Agent Atlas series. The architecture is the context; the harness is the subject. The harness is what keeps the agent inside the layers.

## Run it

```bash
npm install
npm run verify
npm run dev
```

The server listens on port 3000. Set `PORT` to use another port.

## Layers

| Folder              | What goes here                                  | May import               |
| ------------------- | ----------------------------------------------- | ------------------------ |
| `src/models/`       | Entity types and `AppError`                     | nothing                  |
| `src/repositories/` | In-memory storage, one class per entity         | `models`                 |
| `src/services/`     | Business rules                                  | `repositories`, `models` |
| `src/controllers/`  | Express routers: validate input, call a service | `services`, `models`     |
| `src/app.ts`        | Composition root: builds every layer            | every layer              |

The rule: controllers call services, services call repositories, and nothing goes back up. The full description is in [`AGENTS.md`](AGENTS.md).

## What the harness adds to starting-point

- `AGENTS.md`, with the layers table and the verification commands. It is an open format that several coding agents read from the project root.
- The skill `.agents/skills/add-endpoint/SKILL.md`, which lists the steps for a new endpoint in layer order. It uses the `SKILL.md` format of the Agent Skills standard.
- Architecture rules in `.dependency-cruiser.cjs`, run by `npm run check:architecture`.
- A stop hook, `.agents/hooks/check-architecture-before-stop.mjs`. When the agent tries to finish, it runs the architecture check. If it fails, it prints the output to stderr and exits with code 2, so the agent cannot finish yet.
- The [security piece](../security/): `.agents/permissions.json` (the agent never reads `.env` or `.env.*`, and asks before `git push` or installing a dependency), a hook that blocks dangerous shell commands (`.agents/hooks/block-dangerous-commands.mjs`), and a stop hook that runs the secret scanner secretlint (`.agents/hooks/check-secrets-before-stop.mjs`, `.secretlintrc.json`, `npm run check:secrets`). Its tests are `test/block-dangerous-commands.test.ts` and `test/check-secrets-before-stop.test.ts`.
- The [quality piece](../quality/): a hook that formats each file the agent edits (`.agents/hooks/format-after-edit.mjs`), a stop hook that runs the formatting check, ESLint, TypeScript and the tests (`.agents/hooks/check-quality-before-stop.mjs`), and the ESLint and Prettier settings (`eslint.config.js`, `prettier.config.mjs`, `.prettierignore`). ESLint limits the size of a function and ignores `eslint-disable` comments. Its tests are `test/format-after-edit.test.ts` and `test/check-quality-before-stop.test.ts`.
- The [clean-code piece](../clean-code/): the duplicate detector jscpd (`.jscpd.json`, `npm run check:duplicates`), a stop hook that reports each new duplicate once as a warning (`.agents/hooks/report-duplicates-before-stop.mjs`), a review skill (`.agents/skills/clean-code-review/SKILL.md`) and the `Clean code` section of `AGENTS.md`. Its tests are in `test/report-duplicates-before-stop.test.ts`.
- A service-level test file, `test/loan-service.test.ts`, that checks the loan rules without HTTP.

## Connect it to your agent

The harness is not tied to one agent. Agents do not share a settings format, so the logic lives in `.agents/` and each agent's own settings point to it. Until you connect them, the files in `.agents/` do nothing on their own.

- **Stop hooks:** register `node .agents/hooks/check-architecture-before-stop.mjs`, `node .agents/hooks/check-secrets-before-stop.mjs`, `node .agents/hooks/check-quality-before-stop.mjs` and `node .agents/hooks/report-duplicates-before-stop.mjs` as your agent's stop hooks, the event that runs when the agent is about to finish. The scripts find the project root from their own location, so they work from any current directory. Your agent must read exit code 2, with the reason on stderr, as "not done yet". If it expects another output format, it needs a small adapter.
- **Before a shell command:** register `node .agents/hooks/block-dangerous-commands.mjs` on the event that runs before a tool or shell command. Exit code 2 means "do not run it".
- **After an edit:** register `node .agents/hooks/format-after-edit.mjs` on the event that runs after the agent writes or edits a file.
- **Skills:** some agents read `.agents/skills/` directly. For the others, copy `.agents/skills/add-endpoint/` and `.agents/skills/clean-code-review/` to the folder where your agent looks for skills.
- **Permissions:** copy the rules in `.agents/permissions.json` into your agent's permission settings.

The Agent Atlas page for this harness shows the settings for some agents and says what was tested: <https://agentatlas-dev.vercel.app/en/harnesses/layered/>.

## Checks

| Command                      | What it checks                                                                                                           | Run by a stop hook     |
| ---------------------------- | ------------------------------------------------------------------------------------------------------------------------ | ---------------------- |
| `npm test`                   | Acceptance tests (11), service tests (3), security hook tests (39), quality hook tests (12) and duplicate hook tests (5) | yes (quality)          |
| `npm run check:architecture` | Layer rules                                                                                                              | yes (architecture)     |
| `npm run check:secrets`      | Secrets in the files that `.gitignore` keeps                                                                             | yes (secrets)          |
| `npm run typecheck`          | TypeScript types                                                                                                         | yes (quality)          |
| `npm run lint`               | ESLint                                                                                                                   | yes (quality)          |
| `npm run format:check`       | Prettier formatting                                                                                                      | yes (quality)          |
| `npm run verify`             | All of the above, in order                                                                                               | no                     |
| `npm run check:duplicates`   | Duplicated code in `src/` (a report, it never fails)                                                                     | yes (duplicates, once) |

## Limits

- `.agents/permissions.json` only lists the rules. How far they reach depends on the agent's permission settings: they usually cover the agent's file tools, not a script that opens the file itself. The dangerous command hook reads the text of a command, so a script that deletes files or reads `.env` from inside a program also gets past it. Only a sandbox stops that. The limits of the security piece are in [its README](../security/README.md).
- The stop hooks run on every stop. The quality hook is the slowest, because it runs the linter, the type checker and every test. A slow test suite makes every stop slow. The limits of the quality piece are in [its README](../quality/README.md).
- The duplicate hook only finds copied code, not the same logic written with other names, and it cannot tell a duplicate that should be one function from code that only looks alike. Names, responsibilities and scope are left to the `Clean code` rules and the review skill. The limits of the clean-code piece are in [its README](../clean-code/README.md).
- How many times in a row a stop hook can send the agent back depends on the agent. The scripts do not limit it.

The full contract (entities, status codes, error codes and acceptance tests) is in [`../MINI_APP_SPEC.md`](../MINI_APP_SPEC.md).
{
  "compilerOptions": {
    "strict": true,
    "target": "ES2022",
    "module": "ESNext",
    "moduleResolution": "Bundler",
    "noEmit": true,
    "types": ["node"],
    "skipLibCheck": true,
    "noUncheckedIndexedAccess": true,
    "verbatimModuleSyntax": true
  },
  "include": ["src", "test"]
}

How to use it

You can run it in the browser or on your computer.

In the browser, the Run in the browser button opens the harness in StackBlitz, an external service, in a new tab. StackBlitz installs the dependencies and runs npm test, which here is 70 tests. On October 18, 2026, with the clean-code piece added, I tested it in Chromium 152 with the “Open StackBlitz in this tab” button: all 70 tests passed, and its terminal measured 56 s to install the dependencies and run them. Before that, I had to give the format and command hook tests more time: that day they went over Vitest’s 5-second limit, also in the version published on October 16. I did not test it in Firefox or Safari, or with the new-tab button.

On your computer you need Node 22 or later. Copy the harness files into a folder and run these commands inside it, with npm and not pnpm:

npm install
npm test
npm run verify
npm run dev

npm test runs the 70 tests: the 11 acceptance tests, 3 for the loan service, 39 for the security hooks, 12 for the quality hooks, and 5 for the duplicate hook. npm run verify runs everything AGENTS.md asks for: types, lint, formatting, the layer rules, the secret scanner, and the tests. npm run check:duplicates shows the duplicated code in src/ and never fails. npm run dev starts the API at http://localhost:3000.

Connect it to your agent

Each agent reads hooks, skills, and permissions from its own settings, and there is no common format. That is why the harness keeps the logic in .agents/, and each agent only needs a few settings files that point there. Those files are not part of the harness: create them yourself in the harness folder, or in the root of the project where you copied it.

The hook scripts do not depend on the agent. They find the project root from their own location and, if something fails, exit with code 2 and write the reason to stderr (the error output). According to their docs, Claude Code and Codex understand that code in the stop hook, in the hook that runs before a command, and in the hook that runs after a file edit.

Claude Code. Create .claude/settings.json with the hooks and the permission rules:

{
  "permissions": {
    "deny": ["Read(./.env)", "Read(./.env.*)"],
    "ask": ["Bash(git push *)", "Bash(npm install *)", "Bash(npm i *)", "Bash(npm add *)"]
  },
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash|PowerShell",
        "hooks": [
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/block-dangerous-commands.mjs"],
            "timeout": 30
          }
        ]
      }
    ],
    "PostToolUse": [
      {
        "matcher": "Edit|Write",
        "hooks": [
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/format-after-edit.mjs"],
            "timeout": 30
          }
        ]
      }
    ],
    "Stop": [
      {
        "hooks": [
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/check-architecture-before-stop.mjs"],
            "timeout": 120
          },
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/check-secrets-before-stop.mjs"],
            "timeout": 120
          },
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/check-quality-before-stop.mjs"],
            "timeout": 600
          },
          {
            "type": "command",
            "command": "node",
            "args": ["${CLAUDE_PROJECT_DIR}/.agents/hooks/report-duplicates-before-stop.mjs"],
            "timeout": 120
          }
        ]
      }
    ]
  }
}

The command with args form needs Claude Code 2.1.139 or later. Also:

  • Claude Code reads AGENTS.md on its own since version 2.1.277, when there is no CLAUDE.md. With an earlier version, create a CLAUDE.md with a single line, @AGENTS.md, so it loads.
  • Claude Code reads nothing inside .agents/. To get the skills, copy the .agents/skills/add-endpoint/ and .agents/skills/clean-code-review/ folders into .claude/skills/.

Codex. It reads AGENTS.md and the skills in .agents/skills/ on its own. For the hooks, create .codex/hooks.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "^Bash$",
        "hooks": [
          {
            "type": "command",
            "command": "node .agents/hooks/block-dangerous-commands.mjs",
            "timeout": 30
          }
        ]
      }
    ],
    "PostToolUse": [
      {
        "matcher": "^apply_patch$",
        "hooks": [
          {
            "type": "command",
            "command": "node .agents/hooks/format-after-edit.mjs",
            "timeout": 30
          }
        ]
      }
    ],
    "Stop": [
      {
        "hooks": [
          {
            "type": "command",
            "command": "node .agents/hooks/check-architecture-before-stop.mjs",
            "timeout": 120
          },
          {
            "type": "command",
            "command": "node .agents/hooks/check-secrets-before-stop.mjs",
            "timeout": 120
          },
          {
            "type": "command",
            "command": "node .agents/hooks/check-quality-before-stop.mjs",
            "timeout": 600
          },
          {
            "type": "command",
            "command": "node .agents/hooks/report-duplicates-before-stop.mjs",
            "timeout": 120
          }
        ]
      }
    ]
  }
}

Codex runs the command from the folder where you started the session, so start it in the harness folder. Before the hooks run, you have to review them and mark them as trusted with /hooks. The rules that block reading .env go in a permission profile in its config.toml, a feature its docs mark as beta. To make it ask before git push or installing, Codex uses rules in .codex/rules/; the security piece has an example. Its docs also say that, at the stop, it expects JSON on standard output when the hook exits with code 0. The scripts write nothing in that case, and I could not check whether Codex accepts that.

Cursor. It reads AGENTS.md and the skills in .agents/skills/. Its stop hook expects JSON with followup_message on standard output, and its docs do not say what it does with code 2 at that moment. The script does not use that format, so it would need a small adapter, which I did not write. For .env, Cursor uses the .cursorignore file; according to its docs, it does not stop the agent from reading the file with terminal commands. The command hook does connect to its beforeShellExecution event, and the format hook to afterFileEdit; the security and quality pieces explain how and what is still unconfirmed.

Other agents. If your agent’s stop hook understands code 2 with the reason on stderr, the script works without changes. If it expects another format, it needs an adapter.

I did not test any of these setups in a real agent session: they come from each agent’s docs, accessed on October 14, 2026, on October 15 for the security part, on October 16 for the quality part, and on October 18 for the clean-code part. I did test the script with a simulated stop event on stdin, the way these agents send it; that is in the next section.

What the hook checks

This part is about the layers hook. The other hooks are explained in the security piece, the quality piece, and the clean-code piece.

The hook is the script .agents/hooks/check-architecture-before-stop.mjs. Once it is connected, the agent runs it every time it tries to finish. The script runs npm run check:architecture, which checks the layer rules. If it fails, it writes the output to stderr and exits with code 2. In Claude Code and Codex, according to their docs, that means the agent cannot finish yet: the message goes back to the agent, which has to fix the code.

There are two things the hook does not do:

  • It does not run the tests, the linter, the formatting check, or the type check. Until October 16 it also ran the tests; now all of that is done by the hook of the quality piece, which connects to the same stop.
  • The script does not read the stop_hook_active field in its input, so it does not limit how many times in a row it sends the agent back to work. That limit depends on the agent. According to its docs, Claude Code stops after 8 consecutive continuations, and the count resets when the agent uses a tool. The Codex docs do not describe a limit: if the tests fail for a reason the agent cannot fix, you will have to interrupt the session.

What I checked on October 14, 2026, on Windows 11 with Node 24.19, when the hook was called check-before-stop.mjs and also ran the tests:

  • npm run verify passed: types, lint, format, no dependency violations found (15 modules, 39 dependencies cruised), and 14 of 14 tests.
  • I broke a rule on purpose: I added import { BookRepository } from '../repositories/book-repository'; to books-controller.ts. npm run check:architecture failed with error controllers-do-not-import-repositories: src/controllers/books-controller.ts → src/repositories/book-repository.ts. I ran the hook with a simulated Stop event on stdin: it exited with code 2 and printed Stop check failed. Fix these before you finish.. Then I undid the change.
  • I changed the loan limit from 3 to 4: the hook also exited with code 2, because tests failed. Then I set the limit back to 3.
  • After moving the hook to .agents/, I repeated the test and started it from another folder, as an agent that starts somewhere else would. With the correct code it exited with code 0 and no output; with the BookRepository import, it exited with code 2 and the same error.
  • The hook took between 3 and 5 seconds on that machine.

On October 15, 2026, after adding the security piece, npm run verify passed again on the same machine, now with the secret scanner and 53 of 53 tests.

On October 16, with the quality piece added, npm run verify passed with 65 of 65 tests. I repeated the BookRepository import in books-controller.ts: the layers hook exited with code 2 in about 2.4 seconds and printed Architecture check failed. With the loan limit at 4, it is now the quality hook that exits with code 2.

On October 18, with the clean-code piece added, npm run verify passed with 70 of 70 tests. jscpd found no duplicated code in src/, so the harness code did not change.

These runs fed the script a simulated Stop event, not a real session of any agent.

Limitations

  • dependency-cruiser checks imports, not where the logic lives. A business rule written inside a controller is not detected, so it has to be reviewed by hand.
  • The files in .agents/ do nothing on their own. The hooks and the permissions only work once you connect them in your agent’s settings, and I did not test that connection in a real session.
  • The stop hooks run on every stop, including when the agent only answers a question. On October 16, on that machine, the layers hook took about 2.4 seconds, the secrets hook about 1.5, and the quality hook about 14.5. On October 18, with the clean-code piece and other work on the machine, the duplicate hook took about a second and the quality hook about 15, one more than without the 5 new tests. The four started at the same time, as Claude Code does, took about 15.
  • The .env deny rule is not a sandbox. In Claude Code, for example, it covers its file tools and the file commands it recognizes in Bash, such as cat, but not a script that opens the file on its own. According to its permissions docs, only operating-system sandboxing covers that case. The command hook of the security piece also blocks cat .env in the terminal, but it is not a sandbox either.
  • Nothing technical stops the agent from changing the hooks, the permissions, or .dependency-cruiser.cjs. AGENTS.md forbids it, but that is an instruction, not a barrier.
  • I only tested it on Windows 11 with Node 24.19. I did not test it on macOS, on Linux, or with Node 22.

Next step

All three pieces are already in this harness: security (day 3), quality (day 4), and clean code (day 5). Piece: security, Piece: quality, and Piece: clean code explain them. The hexagonal harness, from day 6, uses the same pieces and compares its architecture with this one. You can see the full calendar in Harnesses by architecture.

Sources

Accessed on October 14, 2026:

  • Verification and evaluationTechnique

    How to check that the agent did the job right without taking its word for it: give it ways to verify itself, review both the result and the process, and compare setups honestly.

    Published
  • Safe agent executionTechnique

    Why written rules are not enough to contain an agent, and which layers of the environment limit the damage even when the agent makes a mistake or someone tricks it.

    Published