Spaces:
Runtime error
Runtime error
| """Encoding / character issues (10 issues).""" | |
| import numpy as np | |
| import pandas as pd | |
| from .base import Issue, pick_rows | |
| def _sql_injection(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| payloads = [ | |
| "'; DROP TABLE user_events; --", | |
| "1 OR 1=1", | |
| "' UNION SELECT * FROM users --", | |
| "admin'--", | |
| "\" OR \"1\"=\"1", | |
| ] | |
| df.loc[mask, "search_query"] = rng.choice(payloads, size=mask.sum()) | |
| return df | |
| def _html_tags(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "page_url"] = df.loc[mask, "page_url"].apply( | |
| lambda u: f"<b>{u}</b>" if isinstance(u, str) else "<b>url</b>" | |
| ) | |
| return df | |
| def _emoji_in_event_type(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| emojis = ["click🖱️", "purchase🛒", "view👀", "login🔐", "search🔍"] | |
| df.loc[mask, "event_type"] = rng.choice(emojis, size=mask.sum()) | |
| return df | |
| def _null_bytes(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "city"] = df.loc[mask, "city"].apply( | |
| lambda c: c + "\x00" if isinstance(c, str) else "\x00" | |
| ) | |
| return df | |
| def _very_long_string(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "search_query"] = "A" * 10_001 | |
| return df | |
| def _control_characters(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "city"] = df.loc[mask, "city"].apply( | |
| lambda c: c + "\x01\x02\x03" if isinstance(c, str) else "\x01\x02" | |
| ) | |
| return df | |
| def _unicode_special_chars(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| specials = [ | |
| "\u200b", # zero-width space | |
| "\u200f", # right-to-left mark | |
| "\ufeff", # BOM | |
| "\u202e", # right-to-left override | |
| "\u0000", # null | |
| ] | |
| df.loc[mask, "email"] = df.loc[mask, "email"].apply( | |
| lambda e: e + specials[0] if isinstance(e, str) else specials[0] | |
| ) | |
| return df | |
| def _xss_in_url(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| xss = [ | |
| "http://example.com/<script>alert(1)</script>", | |
| "http://example.com/?q=<img src=x onerror=alert(1)>", | |
| "javascript:alert('xss')", | |
| ] | |
| df.loc[mask, "page_url"] = rng.choice(xss, size=mask.sum()) | |
| return df | |
| def _newlines_in_string(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "city"] = df.loc[mask, "city"].apply( | |
| lambda c: c + "\nInjected\nLine" if isinstance(c, str) else "\nInjected" | |
| ) | |
| return df | |
| def _tabs_in_string(df, rate, rng): | |
| df = df.copy() | |
| mask = pick_rows(df, rate, rng) | |
| df.loc[mask, "city"] = df.loc[mask, "city"].apply( | |
| lambda c: c + "\t\tTabbed" if isinstance(c, str) else "\tTabbed" | |
| ) | |
| return df | |
| ISSUES = [ | |
| Issue("sql_injection", "SQL injection in search_query", "encoding", "SQL injection payload in search_query field", "critical", 0.01, _sql_injection), | |
| Issue("html_tags", "HTML tags in page_url", "encoding", "Raw HTML tags injected into page_url field", "high", 0.02, _html_tags), | |
| Issue("emoji_in_event_type", "Emoji characters in event_type", "encoding", "event_type contains emoji characters", "medium", 0.02, _emoji_in_event_type), | |
| Issue("null_bytes", "Null bytes in city field", "encoding", "city contains \\x00 null byte characters", "high", 0.02, _null_bytes), | |
| Issue("very_long_string", "String field > 10,000 characters", "encoding", "search_query contains 10,001+ character string", "medium", 0.01, _very_long_string), | |
| Issue("control_characters", "Control characters in city", "encoding", "city contains ASCII control characters \\x01-\\x1f", "medium", 0.02, _control_characters), | |
| Issue("unicode_special_chars", "Zero-width / invisible unicode in email", "encoding", "email contains invisible unicode chars (BOM, RTL mark)", "medium", 0.02, _unicode_special_chars), | |
| Issue("xss_in_url", "XSS script tag in page_url", "encoding", "page_url contains JavaScript injection payload", "critical", 0.01, _xss_in_url), | |
| Issue("newlines_in_string", "Newline characters in city", "encoding", "city field contains embedded newline characters", "medium", 0.02, _newlines_in_string), | |
| Issue("tabs_in_string", "Tab characters in city", "encoding", "city field contains embedded tab characters", "low", 0.02, _tabs_in_string), | |
| ] | |