mirror of
https://github.com/frankwxu/mobile-pii-discovery-agent.git
synced 2026-02-20 13:40:41 +00:00
6 lines
10 KiB
JSON
6 lines
10 KiB
JSON
{"db_path": "selectedDBs\\I4_CloudTabs.db", "PII_type": "EMAIL", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [], "Total_raw_rows": 0, "Exploration_sql": "SELECT device_name AS possible_email\nFROM cloud_tab_devices\nWHERE device_name REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}'\nUNION ALL\nSELECT title AS possible_email\nFROM cloud_tabs\nWHERE title REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}'\nUNION ALL\nSELECT url AS possible_email\nFROM cloud_tabs\nWHERE url REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}'\nUNION ALL\nSELECT key AS possible_email\nFROM metadata\nWHERE key REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}'\nUNION ALL\nSELECT value AS possible_email\nFROM metadata\nWHERE value REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}'", "Extraction_sql": null, "PII_Prompt": "a unique identifier for a destination to which electronic mail (email) can be sent and received over the internet; examples include jane.doe@example.com, john.smith@provider.net, dev-team@startup.io, and user.name+label@domain.org"}
|
|
{"db_path": "selectedDBs\\I4_CloudTabs.db", "PII_type": "PHONE", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [["https://www.google.com/search?q=snack+with+dairy+kids&client=safari&sca_esv=6eac42e19cef7bdf&hl=en-us&sxsrf=ADLYWIJQfVNS-z6971OqeDnskNzgmBnxVA%3A1723546457188&ei=WTu7Zu6YC9vl5NoPsZT5uAM&oq=snack+with+dairy+kids&gs_lp=EhNtb2JpbGUtZ3dzLXdpei1zZXJwIhVzbmFjayB3aXRoIGRhaXJ5IGtpZHNIkXVQmR9Yym5wEXgBkAEAmAGkAaABthWqAQQzMi4yuAEDyAEA-AEBmAIVoALyC6gCD8ICChAAGLADGNYEGEfCAg0QABiABBiwAxhDGIoFwgIHECMYJxjqAsICChAuGIAEGEMYigXCAgoQABiABBhDGIoFwgIQEC4YgAQY0QMYQxjHARiKBcICDhAuGIAEGLEDGNEDGMcBwgILEC4YgAQY0QMYxwHCAgUQABiABMICDhAAGIAEGLEDGIMBGIoFwgILEC4YgAQYkQIYigXCAgsQABiABBiRAhiKBcICDRAAGIAEGEMY5QQYigXCAggQABiABBixA8ICDRAAGIAEGLEDGEMYigXCAhYQLhiABBixAxjRAxhDGIMBGMcBGIoFwgILEAAYgAQYsQMYgwHCAggQLhiABBixA8ICExAuGIAEGLEDGNEDGIMBGMcBGArCAgoQABiABBgUGIcCwgIHEAAYgAQYCsICCBAAGIAEGMkDwgILEC4YgAQYsQMYgwHCAgYQABgWGB6YAwiIBgGQBhGSBwQxOC4zoAfzbQ&sclient=mobile-gws-wiz-serp"], ["https://www.bing.com/search?q=figs+tre+wont+ripen+green&form=QBLH&sp=-1&ghc=1&lq=0&pq=figs+tre+wont+ripen+green&sc=10-25&qs=n&sk=&cvid=E8428067064D4726B92401B41C5C64A5&ghsh=0&ghacc=0&ghpl="]], "Total_raw_rows": 2, "Exploration_sql": "SELECT device_name AS possible_phone\nFROM cloud_tab_devices\nWHERE device_name REGEXP '\\+?1?[- .]?\\(?[2-9][0-9]{2}\\)?[- .]?[2-9][0-9]{2}[- .]?[0-9]{4}'\n\nUNION ALL\n\nSELECT title AS possible_phone\nFROM cloud_tabs\nWHERE title REGEXP '\\+?1?[- .]?\\(?[2-9][0-9]{2}\\)?[- .]?[2-9][0-9]{2}[- .]?[0-9]{4}'\n\nUNION ALL\n\nSELECT url AS possible_phone\nFROM cloud_tabs\nWHERE url REGEXP '\\+?1?[- .]?\\(?[2-9][0-9]{2}\\)?[- .]?[2-9][0-9]{2}[- .]?[0-9]{4}'\n\nUNION ALL\n\nSELECT key AS possible_phone\nFROM metadata\nWHERE key REGEXP '\\+?1?[- .]?\\(?[2-9][0-9]{2}\\)?[- .]?[2-9][0-9]{2}[- .]?[0-9]{4}'\n\nUNION ALL\n\nSELECT value AS possible_phone\nFROM metadata\nWHERE value REGEXP '\\+?1?[- .]?\\(?[2-9][0-9]{2}\\)?[- .]?[2-9][0-9]{2}[- .]?[0-9]{4}'", "Extraction_sql": null, "PII_Prompt": "a US phone number is a 10-digit NANP number (area code + exchange + line) that may be written as 2023133725, 202-313-3725, (202) 313-3725, 202.313.3725, +1 202 313 3725, or 1-202-313-3725"}
|
|
{"db_path": "selectedDBs\\I4_CloudTabs.db", "PII_type": "USERNAME", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [["falafel hardboiled egg israel - Google Search"], ["Dickens Cider is now Dickins Cider Company | Hard Cider Drinks | Dickens Cider"], ["Blue Bite | Connecting Possibility"], ["Home page - Mapat"], ["snack with dairy kids - Google Search"], ["figs growing green not ripe - Google Search"], ["Hours: Amusement Park & Hershey Attractions | Hersheypark"], ["How to Change a Light Bulb"], ["Tolls"], ["ravensmanorexperience.com"], ["https://www.google.com/search?q=falafel+hardboiled+egg+israel&ie=UTF-8&oe=UTF-8&hl=en-us&client=safari#ip=1"], ["https://www.dickinscidercompany.com/"], ["https://www.bluebite.com/?utm_source=https://www.themtag.com/n10053654&utm_medium=301"], ["https://www.mapat.org.il/eng/main"], ["https://www.google.com/search?q=snack+with+dairy+kids&client=safari&sca_esv=6eac42e19cef7bdf&hl=en-us&sxsrf=ADLYWIJQfVNS-z6971OqeDnskNzgmBnxVA%3A1723546457188&ei=WTu7Zu6YC9vl5NoPsZT5uAM&oq=snack+with+dairy+kids&gs_lp=EhNtb2JpbGUtZ3dzLXdpei1zZXJwIhVzbmFjayB3aXRoIGRhaXJ5IGtpZHNIkXVQmR9Yym5wEXgBkAEAmAGkAaABthWqAQQzMi4yuAEDyAEA-AEBmAIVoALyC6gCD8ICChAAGLADGNYEGEfCAg0QABiABBiwAxhDGIoFwgIHECMYJxjqAsICChAuGIAEGEMYigXCAgoQABiABBhDGIoFwgIQEC4YgAQY0QMYQxjHARiKBcICDhAuGIAEGLEDGNEDGMcBwgILEC4YgAQY0QMYxwHCAgUQABiABMICDhAAGIAEGLEDGIMBGIoFwgILEC4YgAQYkQIYigXCAgsQABiABBiRAhiKBcICDRAAGIAEGEMY5QQYigXCAggQABiABBixA8ICDRAAGIAEGLEDGEMYigXCAhYQLhiABBixAxjRAxhDGIMBGMcBGIoFwgILEAAYgAQYsQMYgwHCAggQLhiABBixA8ICExAuGIAEGLEDGNEDGIMBGMcBGArCAgoQABiABBgUGIcCwgIHEAAYgAQYCsICCBAAGIAEGMkDwgILEC4YgAQYsQMYgwHCAgYQABgWGB6YAwiIBgGQBhGSBwQxOC4zoAfzbQ&sclient=mobile-gws-wiz-serp"], ["https://www.google.com/search?q=figs+growing+green+not+ripe&ie=UTF-8&oe=UTF-8&hl=en-us&client=safari"], ["https://www.hersheypark.com/info/hours/"], ["https://www.angi.com/articles/how-change-lightbulb.htm"], ["https://www.panynj.gov/bridges-tunnels/en/tolls.html"], ["https://www.ravensmanorexperience.com/uploads/b/10bcaef0-743a-11ea-b3f4-31522c100dea/Summer%20PDF_NTkzNT.pdf"], ["Cellebrite"], ["server_change_token"], ["use_manatee_container"], [{"__bytes_b64__": "YnBsaXN0MDDUAQIDBAUGBwpYJHZlcnNpb25ZJGFyY2hpdmVyVCR0b3BYJG9iamVjdHMSAAGGoF8QD05TS2V5ZWRBcmNoaXZlctEICVRyb290gAGkCwwRElUkbnVsbNINDg8QViRjbGFzc18QD0NoYW5nZVRva2VuRGF0YYADgAJPEC8fChASDAAATBJW9UHWAAAAABgBGAAiFgj5m+me89ruvc4BEM2Hk5/Ph+jnpgEoANITFBUWWiRjbGFzc25hbWVYJGNsYXNzZXNfEBNDS1NlcnZlckNoYW5nZVRva2VuohUXWE5TT2JqZWN0CBEaJCkyN0lMUVNYXmNqfH6AsrfCy+HkAAAAAAAAAQEAAAAAAAAAGAAAAAAAAAAAAAAAAAAAAO0="}]], "Total_raw_rows": 24, "Exploration_sql": "SELECT title AS possible_username\nFROM cloud_tabs\nWHERE title REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT url AS possible_username\nFROM cloud_tabs\nWHERE url REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT device_name AS possible_username\nFROM cloud_tab_devices\nWHERE device_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT key AS possible_username\nFROM metadata\nWHERE key REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT value AS possible_username\nFROM metadata\nWHERE value REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b';", "Extraction_sql": null, "PII_Prompt": "a username is a short textual identifier chosen by a user to represent their account or public handle within an application or service it is stored as plain text contains no whitespace does not include a domain component and is intended for human recognition rather than internal system uniqueness"}
|
|
{"db_path": "selectedDBs\\I4_CloudTabs.db", "PII_type": "PERSON_NAME", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [["Cellebrite"], ["falafel hardboiled egg israel - Google Search"], ["Dickens Cider is now Dickins Cider Company | Hard Cider Drinks | Dickens Cider"], ["Blue Bite | Connecting Possibility"], ["Home page - Mapat"], ["snack with dairy kids - Google Search"], ["figs growing green not ripe - Google Search"], ["Hours: Amusement Park & Hershey Attractions | Hersheypark"], ["How to Change a Light Bulb"], ["Tolls"], ["ravensmanorexperience.com"]], "Total_raw_rows": 11, "Exploration_sql": "SELECT device_name AS possible_name\nFROM cloud_tab_devices\nWHERE device_name REGEXP '[A-Za-z][A-Za-z\\\\s\\\\.\\\\-]{1,50}'\n\nUNION ALL\n\nSELECT title AS possible_name\nFROM cloud_tabs\nWHERE title REGEXP '[A-Za-z][A-Za-z\\\\s\\\\.\\\\-]{1,50}'", "Extraction_sql": null, "PII_Prompt": "a loosely structured human name-like strings that typically consist of a first name, a first name and a last name, and may also include middle names, initials, prefixes (e.g., Mr., Dr.), and suffixes (e.g., Jr., Sr.)"}
|
|
{"db_path": "selectedDBs\\I4_CloudTabs.db", "PII_type": "POSTAL_ADDRESS", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [["cool dry place storage - Google Search"], ["Menu | Raven's Manor"], ["https://www.google.com/search?q=cool+dry+place+storage&ie=UTF-8&oe=UTF-8&hl=en-us&client=safari"], ["https://www.bing.com/search?q=figs+tre+wont+ripen+green&form=QBLH&sp=-1&ghc=1&lq=0&pq=figs+tre+wont+ripen+green&sc=10-25&qs=n&sk=&cvid=E8428067064D4726B92401B41C5C64A5&ghsh=0&ghacc=0&ghpl="]], "Total_raw_rows": 4, "Exploration_sql": "SELECT device_name AS possible_address\nFROM cloud_tab_devices\nWHERE device_name REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b'\nUNION ALL\nSELECT title AS possible_address\nFROM cloud_tabs\nWHERE title REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b'\nUNION ALL\nSELECT url AS possible_address\nFROM cloud_tabs\nWHERE url REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b'\nUNION ALL\nSELECT value AS possible_address\nFROM metadata\nWHERE value REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b';", "Extraction_sql": null, "PII_Prompt": "a US postal address is a street-level mailing location in the United States, commonly appearing as a street name and suffix (e.g., 'Market St') optionally with a street number (e.g., '1500 Market St'), unit, city/state, ZIP, or a PO Box (e.g., 'P.O. Box 123')"}
|