Personal Cloud & Photo Storage · Mobile App Developers

Detect Explicit Passcode Disclosures in Photo Vault Hints

Users setting up encrypted photo vault albums frequently type their actual secret passcode directly into the password hint field. This example helps developers evaluate whether a submitted password hint explicitly reveals the plain password or secret answer itself. Application code can then block risky hints and prompt the user to write a safer prompt.

Updated · Examples verified with jev-1.13.0

hintpasswordpersonal cloudscreening

Interactive demo

Vault Security Lock & Risk Gauge

A virtual vault door tests album recovery hints in real time. Hints disclosing raw passcodes trigger a glowing red lock icon and sweep a security risk gauge into the danger zone.

The visitor types or chooses a password hint string in a simulated album creation form. Each run sends one live request to the public API; results are model judgments and can vary.

Verified sample results (2 inputs, jev-1.13.0)
  • Hint explicitly discloses a numeric 4-digit PIN code in text.

    {
      "state": {
        "hint_text": "The passcode is 4892"
      },
      "answers": {
        "risk_score": {
          "type": "score",
          "score": 2,
          "legend": {
            "0": "Safe abstract memory prompt or clue",
            "1": "Suggestive phrase that narrows down possibilities",
            "2": "Explicit disclosure of the exact password or pin"
          },
          "confidence": 1,
          "probabilities": {
            "0": 0,
            "1": 0,
            "2": 1
          }
        },
        "is_explicit_leak": {
          "noul": 0.99,
          "type": "noul"
        }
      }
    }
  • Direct disclosure of an alphanumeric album passcode.

    {
      "state": {
        "hint_text": "Password is secret123"
      },
      "answers": {
        "risk_score": {
          "type": "score",
          "score": 2,
          "legend": {
            "0": "Safe abstract memory prompt or clue",
            "1": "Suggestive phrase that narrows down possibilities",
            "2": "Explicit disclosure of the exact password or pin"
          },
          "confidence": 1,
          "probabilities": {
            "0": 0,
            "1": 0,
            "2": 1
          }
        },
        "is_explicit_leak": {
          "noul": 0.99,
          "type": "noul"
        }
      }
    }

The problem

The input consists of a user-written text string for an album recovery hint, along with the stored secret password. Users often inadvertently paste or type the plain secret directly into the hint, exposing it on the device unlock screen. The application must analyze the hint text to detect explicit disclosures before saving it. Application code uses this triage judgment to show a security warning or allow the hint to be saved.

How TypeSafe helps

In these observed examples, the input state supplies a secret_passcode and hint_text evaluated by the explicit_passcode_disclosure Noul question. The API returns a Noul probability reflecting whether the passcode is explicitly stated in the hint, yielding observed scores of 0.99 when the passcode is typed directly ('My passcode is 8492'), 0.06 for a conceptual hint ('Year my grandmother was born'), and 0.03 for an empty hint string. Application code can consume this probability score to trigger a security warning prompting the user to rephrase their hint before saving. While the API evaluates text semantics to generate the score, application code independently decides threshold policies and manages the unlock hint update workflow.

Input
A short text string entered by the user in the album password hint input field.
Decision
Noul: Does the password hint text explicitly reveal the plain password or secret answer itself?
Next action
Vault app displays a security warning asking the user to rephrase the hint before saving.

Three verified examples

These responses were returned during a previous API verification. New probabilities can differ.

Hint explicitly stating the passcode (primary)

Expected behavior: The hint text explicitly contains the secret passcode '8492', producing a high probability for explicit disclosure.

Input

{
  "hint_text": "My passcode is 8492",
  "secret_passcode": "8492"
}

Previous verification response

jev-1.13.0 ·

{
  "model": "jev-1.13.0",
  "usage": {
    "input_tokens": 354,
    "output_tokens": 25
  },
  "answers": {
    "explicit_passcode_disclosure": {
      "noul": 0.99,
      "type": "noul"
    }
  }
}
Safe conceptual hint without disclosing passcode (alternative)

Expected behavior: The hint text gives a conceptual memory prompt without disclosing the passcode '8492', resulting in a low probability of explicit disclosure.

Input

{
  "hint_text": "Year my grandmother was born",
  "secret_passcode": "8492"
}

Previous verification response

jev-1.13.0 ·

{
  "model": "jev-1.13.0",
  "usage": {
    "input_tokens": 350,
    "output_tokens": 25
  },
  "answers": {
    "explicit_passcode_disclosure": {
      "noul": 0.06,
      "type": "noul"
    }
  }
}
Empty hint text string (edge case)

Expected behavior: The hint text is empty and does not contain or state the passcode '8492', resulting in a low probability of explicit disclosure.

Input

{
  "hint_text": "",
  "secret_passcode": "8492"
}

Previous verification response

jev-1.13.0 ·

{
  "model": "jev-1.13.0",
  "usage": {
    "input_tokens": 345,
    "output_tokens": 25
  },
  "answers": {
    "explicit_passcode_disclosure": {
      "noul": 0.03,
      "type": "noul"
    }
  }
}

Try online

Start with a verified example, edit the request, then ask Jev. A live request is sent only when you press Try online.

Expected behavior: The hint text explicitly contains the secret passcode '8492', producing a high probability for explicit disclosure.

Use non-sensitive test data. Live input goes to typesafe.pro and TypeSafe. Anonymous requests use the gateway’s free rate limit.

3Answer
Previous verification

Probability that the answer is yes

99% yes
yes99%
no1%

Near 1 is a strong yes. Near 0 is a strong no. Around 0.5 means uncertainty. Noul has no separate confidence value.

Next step

The hint text explicitly contains the secret passcode '8492', producing a high probability for explicit disclosure.

jev-1.13.0 · verified 2026-09-28 · new probabilities may differ

The current request, in code
import jsonfrom urllib.error import HTTPError, URLErrorfrom urllib.request import Request, urlopenBASE_URL = "https://api.typesafe.pro"payload = {    "model": "jev-latest",    "state": {        "hint_text": "My passcode is 8492",        "secret_passcode": "8492",    },    "questions": {        "explicit_passcode_disclosure": {            "type": "noul",            "criteria": {                "true": "The hint explicitly contains or states the plain passcode or secret answer.",                "false": "The hint provides a prompt or context without explicitly stating the plain secret.",            },            "instructions": "Does the password hint text explicitly reveal the plain passcode or secret answer?",        },    },}request = Request(    f"{BASE_URL}/v1/systemone",    data=json.dumps(payload).encode("utf-8"),    headers={"Content-Type": "application/json"},    method="POST",)try:    with urlopen(request, timeout=20) as response:        result = json.load(response)except HTTPError as error:    raise SystemExit(f"API returned HTTP {error.code}") from errorexcept URLError as error:    raise SystemExit(f"Connection failed: {error.reason}") from errorprint(json.dumps(result["answers"], indent=2))

Limitations and review

  • This request tests a single Noul question focused on explicit passcode disclosure, but this does not prevent you from using separate questions to check indirect security risks or enforce broader content rules in multi-label workflows.
  • The returned numerical values of 0.99, 0.06, and 0.03 are observed sample outputs from specific evaluation runs rather than guaranteed future responses for similar text.
  • Evaluating semantic disclosure provides an initial triage score; you can combine this judgment with secondary application checks, such as exact substring matching, if your security policy requires rigid pattern validation.

Set thresholds against your own examples before relying on an automated decision.

← Explore all TypeSafe use cases