| from __future__ import annotations |
|
|
| import difflib |
| import json |
| from pathlib import Path |
| from typing import Any |
|
|
| from src.domain import DEFAULT_SEED, RULESET_ID, SCHEMA_ID, Split |
| from src.rules import rule_catalog, ruleset_meta |
| from src.schemas import Scenario |
|
|
| ROOT = Path(__file__).resolve().parents[1] |
|
|
|
|
| def _diff(path: str, before: str, after: str) -> str: |
| old = before.splitlines(keepends=True) |
| new = after.splitlines(keepends=True) |
| if old and not old[-1].endswith("\n"): |
| old[-1] += "\n" |
| if new and not new[-1].endswith("\n"): |
| new[-1] += "\n" |
| lines = list( |
| difflib.unified_diff( |
| old, |
| new, |
| fromfile=f"a/{path}", |
| tofile=f"b/{path}", |
| lineterm="\n", |
| ) |
| ) |
| header = [f"diff --git a/{path} b/{path}\n"] |
| return "".join(header + lines) |
|
|
|
|
| S01_BEFORE = '''"""Lab invoice reader for ledger-api. Synthetic.""" |
| |
| def get_invoice(invoice_id): |
| return {"id": invoice_id, "status": "draft"} |
| ''' |
|
|
| S01_AFTER = '''"""Lab invoice reader for ledger-api. Synthetic.""" |
| |
| def get_invoice(invoice_id): |
| cursor.execute("SELECT * FROM invoices WHERE id = '" + invoice_id + "'") |
| return cursor.fetchone() |
| ''' |
|
|
| S02_BEFORE = '''"""Lab directory listing for filesvc. Synthetic.""" |
| |
| import subprocess |
| |
| def list_dir(name): |
| return name |
| ''' |
|
|
| S02_AFTER = '''"""Lab directory listing for filesvc. Synthetic.""" |
| |
| import subprocess |
| |
| def list_dir(name): |
| return subprocess.check_output("ls " + name, shell=True) |
| ''' |
|
|
| S03_BEFORE = '''"""Lab download helper for filesvc. Synthetic.""" |
| |
| def download(filename): |
| return filename |
| ''' |
|
|
| S03_AFTER = '''"""Lab download helper for filesvc. Synthetic.""" |
| |
| def download(filename): |
| return open("/var/data/" + filename, "rb").read() |
| ''' |
|
|
| S04_BEFORE = '''"""Lab batch import for ledger-api. Synthetic.""" |
| |
| def import_batch(raw_body): |
| return raw_body |
| ''' |
|
|
| S04_AFTER = '''"""Lab batch import for ledger-api. Synthetic.""" |
| |
| import pickle |
| |
| def import_batch(raw_body): |
| return pickle.loads(raw_body) |
| ''' |
|
|
| S05_BEFORE = '''"""Lab URL preview for fetch-proxy. Synthetic.""" |
| |
| def preview(url): |
| return url |
| ''' |
|
|
| S05_AFTER = '''"""Lab URL preview for fetch-proxy. Synthetic.""" |
| |
| import urllib.request |
| |
| def preview(url): |
| return urllib.request.urlopen(url).read() |
| ''' |
|
|
| S06_BEFORE = '''"""Lab greeting for fetch-proxy. Synthetic.""" |
| |
| def greet(name): |
| return name |
| ''' |
|
|
| S06_AFTER = '''"""Lab greeting for fetch-proxy. Synthetic.""" |
| |
| from markupsafe import Markup |
| |
| def greet(name): |
| return Markup("<h1>Hello " + name + "</h1>") |
| ''' |
|
|
| S07_BEFORE = '''"""Lab invoice reader for payments-api. Synthetic.""" |
| |
| def get_invoice(invoice_id): |
| return {"id": invoice_id} |
| ''' |
|
|
| S07_AFTER = '''"""Lab invoice reader for payments-api. Synthetic.""" |
| |
| def get_invoice(invoice_id): |
| cursor.execute("SELECT * FROM invoices WHERE id = %s", (invoice_id,)) |
| return cursor.fetchone() |
| ''' |
|
|
| S08_BEFORE = '''"""Lab audit helper for notes-app. Synthetic.""" |
| |
| import logging |
| |
| logger = logging.getLogger("notes") |
| |
| def audit(username, n): |
| return username |
| ''' |
|
|
| S08_AFTER = '''"""Lab audit helper for notes-app. Synthetic.""" |
| |
| import hashlib |
| import logging |
| |
| logger = logging.getLogger("notes") |
| |
| def audit(username, n): |
| # never call pickle.loads on user input |
| logger.info("audit execute() count=%s user=%s", n, username) |
| return hashlib.md5(b"static-lab-fixture").hexdigest() |
| ''' |
|
|
| S09_BEFORE = '''"""Lab profile view for notes-app. Synthetic.""" |
| |
| def show(user_id): |
| return {"id": user_id} |
| ''' |
|
|
| S09_AFTER = '''"""Lab profile view for notes-app. Synthetic.""" |
| |
| def show(user_id): |
| return db.execute(build_query(user_id)) |
| ''' |
|
|
| S10_BEFORE = '''"""Lab password store for payments-api. Synthetic.""" |
| |
| def store_password(password): |
| return password |
| ''' |
|
|
| S10_AFTER = '''"""Lab password store for payments-api. Synthetic.""" |
| |
| import hashlib |
| |
| def store_password(password): |
| return hashlib.md5(password.encode()).hexdigest() |
| ''' |
|
|
| RELATED = { |
| "ledger-api/db.py": '"""Lab database stub. Cursor is not connected."""\n\ncursor = None\n', |
| "filesvc/paths.py": '"""Lab path constants."""\n\nBASE = "/var/data"\n', |
| "fetch-proxy/net.py": '"""Lab network stub."""\n\nALLOWED_HOSTS = ("example.test",)\n', |
| } |
|
|
| SCENARIOS: list[dict[str, Any]] = [ |
| { |
| "scenario_id": "S01", |
| "family": "sql-injection", |
| "repository": "ledger-api", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Ledger invoice query concatenates the identifier", |
| "title_fa": "پرسوجوی فاکتور دفتر کل شناسه را الحاق میکند", |
| "description": "A Python change in ledger-api builds SQL with string addition. The AST engine should confirm CWE-89 on the added execute line.", |
| "description_fa": "تغییر پایتون در ledger-api دستور SQL را با جمع رشته میسازد. موتور AST باید CWE-89 را روی خط execute تأیید کند.", |
| "path": "ledger-api/invoices.py", |
| "before": S01_BEFORE, |
| "after": S01_AFTER, |
| "related": ["ledger-api/db.py"], |
| "expected": { |
| "findings": [{"family": "sql-injection", "path": "ledger-api/invoices.py", "start_line": 4, "end_line": 4}], |
| "baseline_false_positive": False, |
| }, |
| }, |
| { |
| "scenario_id": "S02", |
| "family": "command-injection", |
| "repository": "filesvc", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Directory listing uses shell=True", |
| "title_fa": "فهرست پوشه با shell=True اجرا میشود", |
| "description": "filesvc starts a shell command from the directory name. This is a confirmed command-injection finding.", |
| "description_fa": "filesvc فرمان پوسته را از نام پوشه میسازد. این یافته تزریق فرمان تأییدشده است.", |
| "path": "filesvc/listing.py", |
| "before": S02_BEFORE, |
| "after": S02_AFTER, |
| "related": ["filesvc/paths.py"], |
| "expected": { |
| "findings": [{"family": "command-injection", "path": "filesvc/listing.py", "start_line": 6, "end_line": 6}], |
| }, |
| }, |
| { |
| "scenario_id": "S03", |
| "family": "path-traversal", |
| "repository": "filesvc", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Download opens a joined user path", |
| "title_fa": "دانلود مسیر الحاقشده کاربر را باز میکند", |
| "description": "open() joins /var/data/ with filename. There is no resolve-and-contain check.", |
| "description_fa": "open مسیر /var/data را با filename الحاق میکند و بررسی محدوده نیست.", |
| "path": "filesvc/download.py", |
| "before": S03_BEFORE, |
| "after": S03_AFTER, |
| "related": ["filesvc/paths.py"], |
| "expected": { |
| "findings": [{"family": "path-traversal", "path": "filesvc/download.py", "start_line": 4, "end_line": 4}], |
| }, |
| }, |
| { |
| "scenario_id": "S04", |
| "family": "insecure-deserialization", |
| "repository": "ledger-api", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Batch import unpickles the request body", |
| "title_fa": "ورود دسته بدنه درخواست را unpickle میکند", |
| "description": "pickle.loads is applied to raw_body. The Space does not execute that call.", |
| "description_fa": "pickle.loads روی raw_body اعمال شده است. این Space آن را اجرا نمیکند.", |
| "path": "ledger-api/batch.py", |
| "before": S04_BEFORE, |
| "after": S04_AFTER, |
| "related": ["ledger-api/db.py"], |
| "expected": { |
| "findings": [{"family": "insecure-deserialization", "path": "ledger-api/batch.py", "start_line": 6, "end_line": 6}], |
| }, |
| }, |
| { |
| "scenario_id": "S05", |
| "family": "ssrf", |
| "repository": "fetch-proxy", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Preview fetches a caller-controlled URL", |
| "title_fa": "پیشنمایش URL کنترلشده را واکشی میکند", |
| "description": "urlopen receives the url argument with no host allowlist.", |
| "description_fa": "urlopen آرگومان url را بدون فهرست میزبان مجاز میگیرد.", |
| "path": "fetch-proxy/preview.py", |
| "before": S05_BEFORE, |
| "after": S05_AFTER, |
| "related": ["fetch-proxy/net.py"], |
| "expected": { |
| "findings": [{"family": "ssrf", "path": "fetch-proxy/preview.py", "start_line": 6, "end_line": 6}], |
| }, |
| }, |
| { |
| "scenario_id": "S06", |
| "family": "xss-template", |
| "repository": "fetch-proxy", |
| "split": Split.DEVELOPMENT.value, |
| "title": "Greeting wraps input in Markup", |
| "title_fa": "سلام ورودی را در Markup میگذارد", |
| "description": "Markup concatenates the name into HTML. Auto-escaping is bypassed.", |
| "description_fa": "Markup نام را داخل HTML میگذارد و escape خودکار را دور میزند.", |
| "path": "fetch-proxy/greet.py", |
| "before": S06_BEFORE, |
| "after": S06_AFTER, |
| "related": ["fetch-proxy/net.py"], |
| "expected": { |
| "findings": [{"family": "xss-template", "path": "fetch-proxy/greet.py", "start_line": 6, "end_line": 6}], |
| }, |
| }, |
| { |
| "scenario_id": "S07", |
| "family": "sql-injection", |
| "repository": "payments-api", |
| "split": Split.TEST.value, |
| "title": "Payments query is parameterized", |
| "title_fa": "پرسوجوی پرداخت پارامتر دارد", |
| "description": "Hard negative: execute uses a constant SQL string and a bound parameter. The AST engine should stay quiet; the keyword baseline still flags execute().", |
| "description_fa": "نمونه منفی سخت: execute رشته ثابت و پارامتر دارد. موتور AST نباید یافته بدهد؛ خط مبنای کلیدواژه هنوز execute را میبیند.", |
| "path": "payments-api/invoices.py", |
| "before": S07_BEFORE, |
| "after": S07_AFTER, |
| "related": [], |
| "expected": { |
| "findings": [], |
| "baseline_false_positive": True, |
| "baseline_families": ["sql-injection"], |
| }, |
| }, |
| { |
| "scenario_id": "S08", |
| "family": "weak-crypto", |
| "repository": "notes-app", |
| "split": Split.TEST.value, |
| "title": "Audit log mentions pickle and execute", |
| "title_fa": "لاگ ممیزی از pickle و execute نام میبرد", |
| "description": "Hard negative: comments and a constant MD5 digest. Keyword baseline fires; AST should not confirm a weakness.", |
| "description_fa": "نمونه منفی سخت: توضیح و چکیده MD5 ثابت. خط مبنای کلیدواژه هشدار میدهد؛ AST ضعف را تأیید نمیکند.", |
| "path": "notes-app/audit.py", |
| "before": S08_BEFORE, |
| "after": S08_AFTER, |
| "related": [], |
| "expected": { |
| "findings": [], |
| "baseline_false_positive": True, |
| "baseline_families": ["sql-injection", "insecure-deserialization", "weak-crypto"], |
| }, |
| }, |
| { |
| "scenario_id": "S09", |
| "family": "sql-injection", |
| "repository": "notes-app", |
| "split": Split.TEST.value, |
| "title": "Profile view calls a missing query builder", |
| "title_fa": "نمای پروفایل سازنده پرسوجوی غایب را صدا میزند", |
| "description": "db.execute(build_query(user_id)) is visible, but build_query is not in the diff or related files. The reviewer must mark insufficient evidence and must not invent a file.", |
| "description_fa": "db.execute(build_query) دیده میشود ولی تعریف build_query در diff یا فایل مرتبط نیست. بازبین باید شواهد ناکافی بگذارد و فایل نسازد.", |
| "path": "notes-app/profile.py", |
| "before": S09_BEFORE, |
| "after": S09_AFTER, |
| "related": [], |
| "expected": { |
| "findings": [], |
| "expect_insufficient": True, |
| "expect_no_patch": True, |
| }, |
| }, |
| { |
| "scenario_id": "S10", |
| "family": "weak-crypto", |
| "repository": "payments-api", |
| "split": Split.TEST.value, |
| "title": "Password stored with MD5", |
| "title_fa": "گذرواژه با MD5 ذخیره میشود", |
| "description": "store_password hashes a password with MD5. Integrity hashes of constants are out of scope; this password helper is in scope.", |
| "description_fa": "store_password گذرواژه را با MD5 چکیده میکند. چکیده صحت مقدار ثابت خارج از محدوده است.", |
| "path": "payments-api/passwords.py", |
| "before": S10_BEFORE, |
| "after": S10_AFTER, |
| "related": [], |
| "expected": { |
| "findings": [{"family": "weak-crypto", "path": "payments-api/passwords.py", "start_line": 6, "end_line": 6}], |
| }, |
| }, |
| ] |
|
|
|
|
| def scenario_records() -> list[Scenario]: |
| rows: list[Scenario] = [] |
| for item in SCENARIOS: |
| rows.append( |
| Scenario( |
| scenario_id=item["scenario_id"], |
| family=item["family"], |
| repository=item["repository"], |
| split=item["split"], |
| title=item["title"], |
| title_fa=item["title_fa"], |
| description=item["description"], |
| description_fa=item["description_fa"], |
| diff_name=f"{item['scenario_id'].lower()}.diff", |
| related_names=list(item["related"]), |
| expected=item["expected"], |
| ) |
| ) |
| return rows |
|
|
|
|
| def scenarios() -> list[Scenario]: |
| return scenario_records() |
|
|
|
|
| def scenario_by_id(scenario_id: str) -> Scenario: |
| for item in scenarios(): |
| if item.scenario_id == scenario_id: |
| return item |
| raise KeyError(scenario_id) |
|
|
|
|
| def scenario_payload(scenario_id: str) -> dict[str, Any]: |
| for item in SCENARIOS: |
| if item["scenario_id"] == scenario_id: |
| return item |
| raise KeyError(scenario_id) |
|
|
|
|
| def load_diff_text(name: str, root: Path | None = None) -> str: |
| folder = (root or ROOT) / "examples" / "diffs" |
| path = folder / name |
| if path.exists(): |
| return path.read_text(encoding="utf-8") |
| scenario_id = name.split(".")[0].upper() |
| payload = scenario_payload(scenario_id) |
| return _diff(payload["path"], payload["before"], payload["after"]) |
|
|
|
|
| def load_related_text(names: list[str]) -> str: |
| parts: list[str] = [] |
| for name in names: |
| text = RELATED.get(name, "") |
| if text: |
| parts.append(f"=== {name} ===\n{text.rstrip()}\n") |
| return "\n".join(parts) |
|
|
|
|
| def write_examples(folder: Path, seed: int = DEFAULT_SEED) -> dict[str, str]: |
| folder.mkdir(parents=True, exist_ok=True) |
| diffs = folder / "diffs" |
| changes = folder / "changes" |
| diffs.mkdir(parents=True, exist_ok=True) |
| changes.mkdir(parents=True, exist_ok=True) |
| records = [] |
| expected: dict[str, Any] = {} |
| splits = {"development": [], "test": [], "seed": seed, "group": "repository_and_family"} |
| for item in SCENARIOS: |
| path = item["path"] |
| diff_text = _diff(path, item["before"], item["after"]) |
| (diffs / f"{item['scenario_id'].lower()}.diff").write_text(diff_text, encoding="utf-8") |
| case = changes / item["scenario_id"].lower() |
| before_dir = case / "before" |
| after_dir = case / "after" |
| related_dir = case / "related" |
| before_dir.mkdir(parents=True, exist_ok=True) |
| after_dir.mkdir(parents=True, exist_ok=True) |
| related_dir.mkdir(parents=True, exist_ok=True) |
| rel = Path(path) |
| (before_dir / rel.name).write_text(item["before"], encoding="utf-8") |
| (after_dir / rel.name).write_text(item["after"], encoding="utf-8") |
| for related_name in item["related"]: |
| related_path = related_dir / Path(related_name).name |
| related_path.write_text(RELATED[related_name], encoding="utf-8") |
| record = scenario_by_id(item["scenario_id"]) |
| records.append(record.model_dump()) |
| expected[item["scenario_id"]] = item["expected"] |
| splits[item["split"]].append( |
| { |
| "scenario_id": item["scenario_id"], |
| "repository": item["repository"], |
| "family": item["family"], |
| } |
| ) |
|
|
| (folder / "scenarios.jsonl").write_text( |
| "\n".join(json.dumps(item, ensure_ascii=False) for item in records) + "\n", |
| encoding="utf-8", |
| ) |
| (folder / "expected.json").write_text(json.dumps(expected, indent=2), encoding="utf-8") |
| (folder / "splits.json").write_text(json.dumps(splits, indent=2), encoding="utf-8") |
| schema = { |
| "schema_id": SCHEMA_ID, |
| "ruleset_id": RULESET_ID, |
| "language": "python", |
| "input": ["unified_diff", "optional_related_files"], |
| "finding": [ |
| "finding_id", |
| "rule_id", |
| "family", |
| "cwe", |
| "path", |
| "start_line", |
| "review_state", |
| "missing_context", |
| "patch", |
| ], |
| "max_changed_lines": 1000, |
| } |
| (folder / "schema.json").write_text(json.dumps(schema, indent=2), encoding="utf-8") |
| (folder / "rules.json").write_text( |
| json.dumps({"meta": ruleset_meta(), "rules": [item.model_dump() for item in rule_catalog()]}, indent=2, default=str), |
| encoding="utf-8", |
| ) |
| return {"folder": str(folder), "n": str(len(SCENARIOS))} |
|
|
