Add test user seed, password validation, and pre-commit security hook

- backend/scripts/seed.py: creates test@example.com on dev startup
- backend/scripts/start_dev.sh: runs migrations + seed + uvicorn --reload
- backend/app/schemas/user.py: password validator (length, case, digit, special char, forbidden words)
- scripts/security_check.py: Docker-based scanner for secrets, dangerous patterns, weak crypto, bandit
- .githooks/pre-commit: runs security_check.py in python:3.12-slim on every commit

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
curo1305
2026-04-12 15:54:23 +02:00
parent 2351b489fe
commit 61cef2eacd
9 changed files with 323 additions and 2 deletions
+32
View File
@@ -0,0 +1,32 @@
#!/bin/sh
# Security pre-commit hook — runs checks inside Docker, no host installs required.
# Install: git config core.hooksPath .githooks
REPO_ROOT="$(git rev-parse --show-toplevel)"
# Collect staged files on the host and pass them into the container as arguments
STAGED=$(git diff --cached --name-only --diff-filter=ACM)
if [ -z "$STAGED" ]; then
echo "[pre-commit] no staged files — skipping security check."
exit 0
fi
echo "[pre-commit] running security checks..."
# Pass staged file list via environment variable
docker run --rm \
-v "$REPO_ROOT":/repo \
-w /repo \
-e STAGED_FILES="$STAGED" \
python:3.12-slim \
sh -c "pip install --quiet bandit && python scripts/security_check.py"
EXIT_CODE=$?
if [ $EXIT_CODE -ne 0 ]; then
echo "[pre-commit] commit blocked by security check."
exit 1
fi
exit 0
+14
View File
@@ -92,6 +92,20 @@ Browser → Vite dev server (:5173)
Always run `git push` immediately after every `git commit`.
## Security hook
A pre-commit hook lives in `.githooks/pre-commit` and runs `scripts/security_check.py` inside a Docker container. It is registered via `git config core.hooksPath .githooks` (already set in this repo).
The hook checks staged files for:
- Hardcoded credentials / secrets
- Dangerous patterns (`eval`, `exec`, `shell=True`, `pickle`)
- Weak cryptography (MD5, SHA1, DES)
- SQL injection risk
- Debug flags left in code
- `bandit` static analysis on all Python files
New clones must run `git config core.hooksPath .githooks` to activate the hook.
## Changelog convention
Every time files are added or modified, append an entry to the relevant file in `changelog/` (one file per date, named `YYYY-MM-DD_<slug>.md`). If a file for today already exists, append to it rather than creating a new one.
+3
View File
@@ -16,6 +16,9 @@ A fullstack SaaS web application built with FastAPI, React, and PostgreSQL.
- Protected dashboard route
- `/api/users/me` — authenticated user info
- 3 separate Docker containers: `db` (PostgreSQL), `backend` (FastAPI), `frontend` (nginx)
- Dev environment seeds a test user automatically on startup (`test@example.com` / `Test123!`)
- Password policy: min 8 chars, upper + lowercase, digit, special character, no common words
- Pre-commit security hook (`scripts/security_check.py`) runs inside Docker on every commit
## Containers
+45 -1
View File
@@ -1,4 +1,43 @@
from pydantic import BaseModel, EmailStr
import re
from pydantic import BaseModel, EmailStr, field_validator
# Common words that must not appear as whole words inside a password.
# Checked case-insensitively with word boundaries.
_FORBIDDEN_WORDS = {
"password", "passwort", "secret", "welcome", "admin", "administrator",
"login", "user", "test", "guest", "master", "dragon", "monkey", "shadow",
"sunshine", "princess", "letmein", "football", "baseball", "soccer",
"hockey", "abc", "qwerty", "keyboard", "computer", "internet", "access",
"hello", "summer", "winter", "spring", "autumn", "flower", "mustang",
"batman", "superman", "donald", "michael", "jessica", "charlie",
}
def _validate_password(v: str) -> str:
errors = []
if len(v) < 8:
errors.append("at least 8 characters")
if not re.search(r"[A-Z]", v):
errors.append("at least one uppercase letter")
if not re.search(r"[a-z]", v):
errors.append("at least one lowercase letter")
if not re.search(r"\d", v):
errors.append("at least one digit")
if not re.search(r'[!@#$%^&*()\-_=+\[\]{};:\'",.<>?/\\|`~]', v):
errors.append("at least one special character")
lower = v.lower()
for word in _FORBIDDEN_WORDS:
# Match the word as a standalone token (surrounded by non-alpha or string boundary)
if re.search(rf"(?<![a-z]){re.escape(word)}(?![a-z])", lower):
errors.append(f'must not contain the word "{word}"')
break
if errors:
raise ValueError("; ".join(errors))
return v
class UserCreate(BaseModel):
@@ -6,6 +45,11 @@ class UserCreate(BaseModel):
password: str
full_name: str | None = None
@field_validator("password")
@classmethod
def password_strength(cls, v: str) -> str:
return _validate_password(v)
class UserOut(BaseModel):
id: str
+34
View File
@@ -0,0 +1,34 @@
"""Create a test user for the dev environment if it doesn't exist yet."""
import asyncio
from sqlalchemy import select
from app.core.security import hash_password
from app.database import AsyncSessionLocal
from app.models.user import User
TEST_EMAIL = "test@example.com"
TEST_PASSWORD = "Test123!"
TEST_NAME = "Test User"
async def seed() -> None:
async with AsyncSessionLocal() as db:
result = await db.execute(select(User).where(User.email == TEST_EMAIL))
if result.scalar_one_or_none():
print(f"[seed] test user already exists: {TEST_EMAIL}")
return
user = User(
email=TEST_EMAIL,
hashed_password=hash_password(TEST_PASSWORD),
full_name=TEST_NAME,
)
db.add(user)
await db.commit()
print(f"[seed] created test user — email: {TEST_EMAIL} pwd: {TEST_PASSWORD}")
if __name__ == "__main__":
asyncio.run(seed())
+11
View File
@@ -0,0 +1,11 @@
#!/bin/sh
set -e
echo "[start] running migrations..."
alembic upgrade head
echo "[start] seeding dev data..."
python scripts/seed.py
echo "[start] starting uvicorn..."
exec uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
@@ -0,0 +1,22 @@
# 2026-04-12 — Test user, password validation, security hook
**Timestamp:** 2026-04-12T14:10:00
## Summary
Added dev seed user, password strength validation, and a Docker-based pre-commit security check hook.
## Files Added
- `backend/scripts/seed.py` — async script that creates `test@example.com / Test123!` if it doesn't exist; safe to run multiple times
- `backend/scripts/start_dev.sh` — dev container entrypoint: runs `alembic upgrade head` → seed → uvicorn --reload
- `scripts/security_check.py` — security scanner: checks staged files for hardcoded secrets, dangerous patterns (eval/exec/shell=True/pickle), weak crypto (MD5/SHA1/DES), SQL injection risk, debug flags; also runs `bandit` on Python files
- `.githooks/pre-commit` — git hook that runs `security_check.py` inside `python:3.12-slim` Docker container; activated via `git config core.hooksPath .githooks`
- `changelog/2026-04-12_security-validation.md` — this file
## Files Modified
- `backend/app/schemas/user.py` — added `_validate_password` with: min 8 chars, uppercase, lowercase, digit, special char, word-boundary check against ~40 forbidden common words; `UserCreate.password_strength` field validator
- `docker-compose.dev.yml` — backend command changed from bare `uvicorn` to `sh scripts/start_dev.sh`
- `CLAUDE.md` — added Security hook section documenting what the hook checks and how to activate it on new clones
- `README.md` — updated Current State to mention test user, password policy, security hook
+1 -1
View File
@@ -4,7 +4,7 @@
services:
backend:
command: uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
command: sh scripts/start_dev.sh
volumes:
- ./backend:/app
+161
View File
@@ -0,0 +1,161 @@
#!/usr/bin/env python3
"""
Security pre-commit checker.
Runs inside a Docker container — do not execute directly on the host.
Checks:
1. Hardcoded secrets / credentials in staged files
2. Dangerous patterns (eval, exec, shell=True, pickle)
3. Weak cryptography (MD5, SHA1 for passwords, DES)
4. SQL injection risk (raw string formatting into queries)
5. Debug/development flags left in code
6. bandit static analysis on Python files
"""
import os
import re
import subprocess
import sys
from pathlib import Path
# ── Patterns ─────────────────────────────────────────────────────────────────
SECRET_PATTERNS = [
# Only match lowercase/camelCase variable names — excludes ALL_CAPS test constants
(r'(?i)(?<![A-Z_])(password|passwd|secret|api_key|apikey|token|auth)(?![A-Z_])\s*=\s*["\'][^"\']{4,}["\']',
"possible hardcoded credential"),
(r'(?i)secret_key\s*=\s*["\'][^"\']{4,}["\'](?!.*change)',
"hardcoded secret_key (use env var)"),
(r'-----BEGIN (RSA |EC |DSA |OPENSSH )?PRIVATE KEY-----',
"private key embedded in source"),
(r'(?i)(aws_access_key_id|aws_secret_access_key)\s*=\s*["\'][A-Za-z0-9/+]{16,}["\']',
"possible AWS credential"),
]
DANGEROUS_PATTERNS = [
(r'\beval\s*\(', "use of eval()"),
(r'\bexec\s*\(', "use of exec()"),
(r'subprocess\.[a-z_]+\(.*shell\s*=\s*True', "subprocess with shell=True"),
(r'pickle\.loads?\(', "use of pickle (deserialization risk)"),
(r'yaml\.load\s*\([^)]*\)(?!\s*#.*safe)', "yaml.load without Loader=yaml.SafeLoader"),
]
WEAK_CRYPTO_PATTERNS = [
(r'\bMD5\b|\bhashlib\.md5\b', "MD5 is not suitable for password hashing"),
(r'\bSHA1\b|\bhashlib\.sha1\b', "SHA1 is not suitable for password hashing"),
(r'\bDES\b(?!C)', "DES is a weak cipher"),
(r'random\.random\(\)|random\.randint\(', "use secrets module for security-sensitive randomness"),
]
SQL_INJECTION_PATTERNS = [
(r'(execute|query)\s*\(\s*[f"\'].*%s.*["\']\.format|f".*SELECT.*{',
"potential SQL injection via string formatting"),
]
DEBUG_PATTERNS = [
(r'\bdebug\s*=\s*True\b', "debug=True left in code"),
(r'print\s*\(.*password', "possible password printed to stdout"),
]
ALL_PATTERNS = (
[("SECRET", p, m) for p, m in SECRET_PATTERNS]
+ [("DANGER", p, m) for p, m in DANGEROUS_PATTERNS]
+ [("CRYPTO", p, m) for p, m in WEAK_CRYPTO_PATTERNS]
+ [("SQLINJ", p, m) for p, m in SQL_INJECTION_PATTERNS]
+ [("DEBUG", p, m) for p, m in DEBUG_PATTERNS]
)
# Files/dirs to skip
SKIP_DIRS = {".git", "node_modules", "__pycache__", ".venv", "dist", "build", "changelog"}
SKIP_FILES = {"package-lock.json", "poetry.lock", "security_check.py"}
# Skip documentation and binary files — they may reference pattern keywords without risk
SKIP_EXTENSIONS = {
".md", ".txt", ".rst",
".png", ".jpg", ".jpeg", ".gif", ".ico",
".woff", ".woff2", ".ttf", ".eot",
}
# ── Helpers ───────────────────────────────────────────────────────────────────
def get_staged_files() -> list[Path]:
"""Read staged files from STAGED_FILES env var (set by the pre-commit hook)."""
raw = os.environ.get("STAGED_FILES", "")
files = []
for line in raw.splitlines():
line = line.strip()
if not line:
continue
p = Path(line)
if p.suffix in SKIP_EXTENSIONS:
continue
if any(part in SKIP_DIRS for part in p.parts):
continue
if p.name in SKIP_FILES:
continue
if p.exists():
files.append(p)
return files
def scan_file(path: Path) -> list[tuple[int, str, str]]:
findings = []
try:
content = path.read_text(errors="ignore")
except Exception:
return findings
for line_no, line in enumerate(content.splitlines(), start=1):
for category, pattern, message in ALL_PATTERNS:
if re.search(pattern, line):
# Skip lines that are clearly comments explaining the pattern
stripped = line.strip()
if stripped.startswith("#") or stripped.startswith("//"):
continue
findings.append((line_no, category, message))
return findings
def run_bandit(py_files: list[Path]) -> tuple[bool, str]:
if not py_files:
return True, ""
result = subprocess.run(
["python", "-m", "bandit", "-q", "-ll", "--", *[str(f) for f in py_files]],
capture_output=True, text=True
)
passed = result.returncode == 0
return passed, result.stdout + result.stderr
# ── Main ──────────────────────────────────────────────────────────────────────
def main() -> int:
staged = get_staged_files()
if not staged:
print("[security] no staged files to check")
return 0
print(f"[security] scanning {len(staged)} staged file(s)...")
violations = 0
for path in staged:
findings = scan_file(path)
for line_no, category, message in findings:
print(f" [{category}] {path}:{line_no}{message}")
violations += 1
py_files = [f for f in staged if f.suffix == ".py"]
bandit_ok, bandit_out = run_bandit(py_files)
if not bandit_ok:
print("\n[security] bandit found issues:")
print(bandit_out)
violations += 1
if violations:
print(f"\n[security] BLOCKED — {violations} issue(s) found. Fix them or use git commit --no-verify to override.")
return 1
print("[security] all checks passed.")
return 0
if __name__ == "__main__":
sys.exit(main())