Privacy Policy

Prompt SecOps, LLC ("Prompt SecOps", "we", "us", or "our") operates Necio, an AI security testing and behavioral compliance platform ("Service"). This Privacy Policy explains how we collect, use, disclose, and protect information about you when you use our Service.

This Privacy Policy is a notice of our data practices. By accessing or using Necio, you acknowledge that you have read and understand this Privacy Policy. If you do not agree with our data practices, please discontinue use of the Service.

AI Can Make Mistakes — Human Verification Required

Necio uses artificial intelligence to generate prompts, grade and verify responses, and produce reports, and that processing operates on data you submit. AI output can be inaccurate, incomplete, or fabricated, and it may mischaracterize a system, a finding, or a person referenced in scan content. The automated detection and masking described below are likewise imperfect and may miss values they do not recognize.

For this reason, Service output is not an authoritative record of any person, system, or event. A qualified human must review and independently verify AI-generated results before you rely on them, share them, or make any decision about an individual based on them. Where you act as the controller of personal data contained in scan prompts, responses, or evidence, satisfying accuracy obligations for that content under applicable data protection law remains your responsibility.

To the maximum extent permitted by applicable law, Prompt SecOps is not responsible or liable for decisions made or actions taken in reliance on AI-generated output. See Sections 8, 9, and 10 of our Terms of Service.

Masking of Sensitive Data Detection & Safe Telemetry

Wherever possible, Necio tries to mask sensitive information it detects in scan evidence, including Social Security numbers, API keys, credentials, financial identifiers, email addresses, phone numbers, and labeled personal information. Whenever a supported pattern is detected, reports, dashboards, exports, notifications, analytics, and operational telemetry are shown masked copies of the detected values rather than the raw data. Masked evidence retains surrounding context and detection categories so you can investigate an exposure without duplicating the detected value in these outputs.

Original prompts and responses remain available to authorized testing and grading workflows under the retention schedule below, including providers used to grade responses and continuity between campaign turns. Masking does not change the grader input, verdict, confidence, or severity. Browser artifacts use a sanitized text reconstruction with sensitive spans crossed out before image creation; source images and other uninspectable media are omitted. Older, unverified screenshots are withheld from display and export. Automated detection can miss unfamiliar, encoded, or context-dependent sensitive data; a result with no detected pattern is not a guarantee that the source contains no personal information.

1. Information We Collect

Account Information

When you register, we collect your email address, name, company name, country, and a hashed password. We do not store plaintext passwords. If you sign in through a third-party identity provider (such as Google), we receive your name, email address, and verified-email status from that provider; we do not receive or store your password for that provider.

Scan & Usage Data

Necio stores prompts, model responses, and grades from security scans and Maestro campaigns you run against your own AI systems. When enabled, it also stores private screenshot evidence. This data is linked to your account for reports and behavioral drift tracking. Raw content and detailed evidence are minimized or deleted automatically under the retention schedule in Section 4. You must have permission to test each target.

The same applies whichever kind of target you configure: an API endpoint you hold a key for, a browser-based chat interface Necio drives with an automated browser, or a bot you reach through a messaging channel. Each of those is described in more detail below, because each one sends your test content somewhere different.

Your role in scan content. Scan prompts and the responses returned by systems you test may incidentally contain personal data belonging to third parties. You are responsible for ensuring you have a lawful basis to submit that content to the Service. Where data protection law applies to such content, you act as the controller (or equivalent) of that data and we process it on your behalf and on your instructions to deliver the Service.

Target Credentials & Access Details

To reach the system you want to test, you may need to give Necio a way in. Depending on the target type, that can be an API key for an LLM provider, sign-in credentials for a chat interface that requires an account (for example, a Microsoft account when you test Copilot), or a bot token, gateway token, or channel identifier for an agent reached through Discord, Slack, Telegram, Microsoft Teams, or an agent gateway URL.

The same rules apply to all these credentials. Credentials for one-time scans are not retained after the run. Scheduled scans store required credentials encrypted at rest so future runs can run unattended. Saved credentials are never displayed or returned to the browser, even in encrypted form. We use them only to connect to the target you configured.

Use a dedicated testing account. Where a target requires you to sign in, we strongly recommend an account created for testing, holding the least privilege needed, rather than a personal account or one with access to production data. Never submit credentials belonging to someone else.

We recommend rotating credentials according to your organization's credential-management policy and immediately after actual or suspected exposure. You can replace the credential for an existing job from the Schedule page by opening Edit schedule and using the API credential update. The replacement is used by future scheduled runs; deleting the schedule removes its saved target credential.

Files You Attach

Tests can include a document or image attached to the prompt. We store uploaded files so the scan can send them to the target. They are automatically deleted 48 hours after upload by a scheduled purge. The file's name, type, and size are recorded alongside the scan result as delivery evidence and follow the ordinary scan-record retention in Section 4.

An attached file is sent to the target you configured, exactly as your prompt is. Attach only files you have the right to submit, and keep personal, confidential, and regulated content in them to what the authorized test actually requires. Necio is not a file store: do not rely on it to hold anything you need to keep.

Browser-Driven Testing

For web chat targets, Necio uses an automated browser session at the URL you provide. A service we operate or contract with runs that session. Your prompts, attachments, and target responses pass through that service. When evidence capture is enabled, masked evidence images also pass through it. The session visits only your configured target. It carries only the credentials you explicitly supplied for that target.

To reach the chat interface, the automated session accepts cookie, consent, and terms-of-use dialogs presented by the target site, including a terms prompt raised by sending a message or attaching a file. You authorize those acceptances when you launch the scan, and you are responsible for whether you are entitled to accept them on that site. Necio does not defeat CAPTCHAs or bot-detection challenges: when a target presents one, or requires a sign-in you have not supplied, the run stops and reports that outcome rather than grading a reply the target never gave.

Necio Assist Browser Extension

Necio Assist is an optional Chrome extension for Assisted Scans. When you connect a chat tab to it, it fills each Assisted Scan prompt into that chat page and, after you send the prompt yourself, reads the chat's reply. It handles only the prompt text it fills in and the text and HTML of that reply, from the one tab you connected. It does not read cookies, passwords, other tabs, or your browsing history, and it does not track clicks or keystrokes.

The extension has no server of its own. It passes the reply to your Necio page in your browser, which sends it to your Necio account to be graded and saved with your scan results, like a reply you paste by hand. It keeps the connected tab, and any reply not yet delivered to your Necio page, in browser session storage, which is cleared when you close the browser. We do not sell this data, transfer it to third parties, or use it for anything other than grading and reporting your scan. The extension can reach only a site you choose, for the current tab when you click its icon or for a site you ask it to remember; you can remove that access, or the extension, at any time in Chrome's extension settings.

Agent Channel Testing

For bots reached through a messaging platform, Necio uses that platform's own API to deliver your prompts. On Discord, Slack, and Microsoft Teams it creates a temporary channel in the workspace or server you configure, invites the bot you name, runs the scan there, and removes the channel afterward. On Telegram it messages the bot directly. Your prompts and the bot's replies transit that messaging platform and are handled under that platform's own data practices and retention, which we do not control and cannot delete on your behalf. Records the platform keeps may outlive the Necio scan record.

Payment Information

Stripe processes subscription payments and PAYGO credit purchases, including authorized automatic top-ups. We do not store full card numbers or card security codes. We receive customer and payment identifiers, payment status, receipt references, and saved-card details such as brand, last four digits, and expiration. We store credit purchases, balances, reservations, deductions, refunds or adjustments, and token and other usage records (including model, feature, timing, and cost) to operate prepaid billing, display usage, prevent duplicate credits, and resolve disputes. For automatic top-up we also retain your selected amount, threshold, monthly limit, and authorization record. Saving a card does not itself authorize automatic top-up.

Web Search Features

Certain features use web search to improve results. When hallucination or factual-accuracy web search is enabled, Necio sends claims extracted from AI responses under review to a third-party search provider so they can be checked against public sources. When prompt generation is grounded in current threat research, Necio sends the name and description of the system you are testing to that provider as a search query. We do not send your API credentials, full scan datasets, or account identity to the search provider. The search provider retrieves the pages and returns text excerpts and web addresses; Necio does not itself fetch, download, or execute content from the sites returned. The search provider processes these queries under its own terms and may retain them; the results returned may include content published by third parties, in any language, and are not reviewed or endorsed by us.

Usage & Analytics Data

We collect usage, diagnostic, and security analytics such as page views, feature usage, navigation paths, interaction events, error events, device and browser information, and IP addresses. We may use session analytics to understand how the Service is used, troubleshoot issues, investigate abuse, preserve security, and resolve disputes. We do not sell this data or use third-party advertising cookies. Sensitive fields such as passwords, API keys, and payment card details are not intended to be captured by analytics tooling.

Cookies & Session Data

We use necessary session cookies to keep you signed in and protect your session. Browser local storage saves interface preferences and unfinished work, such as scan configuration. Product analytics run in memory during each page visit. They do not set analytics cookies or store analytics identifiers in your browser. We do not use third-party advertising cookies or sell your browsing behavior.

Embedded Videos

Some of our public pages include tutorial videos hosted by YouTube, a Google service. Nothing loads from YouTube until you press play. When you do, the video plays in YouTube's privacy-enhanced mode (youtube-nocookie.com), and YouTube receives your IP address and device and browser information and may store cookies or similar data in your browser. YouTube handles that data under Google's own privacy policy, not ours. You can watch the same videos on YouTube directly instead, using the link next to each video.

2. How We Use Your Information

• To provide, operate, and improve the Necio platform
• To execute vulnerability scans and generate security reports on your behalf
• To generate adversarial test prompts and Maestro campaign turns targeted at the system you configured
• To send transactional emails (account verification, password reset, scan notifications)
• To send product updates, announcements, and other marketing about Necio, where permitted by law
• To process billing, maintain PAYGO credit balances, perform authorized automatic top-ups, and manage subscriptions
• To detect, prevent, and respond to fraud, abuse, or security incidents
• To record and evidence your acceptance of our legal terms and required authorizations
• To enforce our agreements and establish, exercise, or defend legal claims
• To comply with legal obligations

We do not use your scan data or credentials to train AI models. How the third-party model providers behind those calls handle your content is described in Section 3.

De-identified and aggregate data. We may create de-identified or aggregated information from Service data, such as vulnerability rates by attack category, model family, or framework control, and use it to operate, secure, and improve the Service, to research AI security trends, and to publish statistics or benchmarks. Aggregate information we use or publish does not identify you, your organization, or the specific systems you test, and does not include the content of your prompts, responses, credentials, or evidence. We maintain de-identified information in de-identified form and do not attempt to re-identify it, except as permitted by law to confirm our de-identification processes work.

Email communications. You can unsubscribe from marketing emails at any time using the link in the email or by contacting us. If you opt out of marketing, we will still send transactional and service messages, such as account security, billing, scan notifications, and legal notices, because they are necessary to provide the Service. We do not currently use open or click tracking in our emails; if we introduce it, we will update this Policy first.

Where the EU or UK General Data Protection Regulation applies, we process personal data on the following legal bases: performance of a contract (providing the Service you signed up for), our legitimate interests (operating, securing, and improving the Service and preventing abuse), compliance with legal obligations, and your consent where required.

Testing controls and abuse investigations. We use account status, scan or campaign identifiers, execution status, timestamps, and relevant authorization records to enforce cancellation, suspension, and platform pauses. We record stop requests and administrative control changes, including the administrator's reason, to investigate misuse and explain execution decisions. These records follow the security and audit retention periods and legal exceptions described below.

3. Data Sharing & Third Parties

We do not sell your personal data. We share data only with:

• Cloud hosting & serverless infrastructure providers: to run the Service and to store security and audit records
• Managed database & storage providers: to store account and scan data
• Payment processors: to process subscription payments, credit purchases, and authorized automatic top-ups
• Product analytics and diagnostic providers: usage, error, and session analytics to operate, secure, and improve the Service
• Video hosting: if you press play on an embedded tutorial video, YouTube (Google) receives the information described under Embedded Videos
• Transactional email delivery providers: to send account and service notifications
• Targets you configure: scan prompts and any attached files are sent to the endpoint, chat interface, or bot you specify
• Browser-automation providers: when you test a browser-based target, your prompts, attachments, the target's replies, and any captured screenshots pass through the automation service that drives the browser session
• Messaging platforms you target: when you test a bot on Discord, Slack, Telegram, or Microsoft Teams, your prompts and its replies transit that platform under its own terms and retention
• AI model providers: scan prompts and responses may be processed by LLM services we operate or contract with to generate test prompts and campaign turns, and to grade, verify, and analyze results (including multi-model verification, where a result may be independently evaluated by more than one model)
• Web search providers: when web-search features are enabled, we send short queries derived from your scan configuration (such as a target system's name and description) or from claims extracted from AI responses under review to a third-party search provider, to retrieve current threat intelligence and to verify factual accuracy
• Integrations you configure: scan results and alerts are delivered to the SIEM endpoints, webhooks, and notification channels you specify
• Internal security notification providers: we may use Telegram and other notification providers for limited operational and security metadata, such as a service identifier, account or run identifier, timestamp, reason code, and action taken. Internal alerts should not include credentials, prompt bodies, model responses, or attached test files

Targets are your choice, not ours. Supported targets include commercial APIs, independent inference hosts, providers outside the United States, self-hosted endpoints, web chats, and messaging bots. You select the target. Your prompts and attachments are sent there and handled under its terms. We do not vet your targets. We cannot retrieve or delete content after a target receives it.

Models Necio itself runs. Separately from the targets you configure, Necio calls AI models of its own to generate test prompts and Maestro campaign turns, to grade and verify responses, and to help produce reports. These calls are routed through a managed AI gateway to the inference providers behind it.

Some of those models are open-weight models, and that is deliberate. Necio uses open-weight models to generate adversarial test prompts and Maestro campaign turns, because attackers use openly available models and a model with fewer built-in refusals writes payloads that resemble what a real adversary would actually send. Testing against realistic payloads is the point of the product. Open-weight models are used only to write the prompts we then send to systems you have authorized us to test. They are not used to grade or verify results, and they are not used to process your account, billing, or credential data.

Retention and training at the model providers. Every model call Necio makes is routed through an AI gateway configured to use only inference providers that have zero-data-retention agreements. Those providers do not retain your scan content or use it to train their models. This applies to the open-weight models that write test prompts and to the models that grade and verify results. It does not cover web search providers, which are described in Section 1. Providers we route to may operate outside the United States. We never send your account credentials, target credentials, provider API keys, or payment data to any of them.

All service providers we contract with are required to process your data only as necessary to deliver their services.

If Prompt SecOps is involved in a merger, acquisition, financing, reorganization, or sale of assets, your information may be transferred as part of that transaction, subject to this Privacy Policy.

We may disclose information if required by law, court order, or to protect the rights, safety, or property of Prompt SecOps, our users, or the public.

4. Data Retention

We retain personal data as needed to provide the Service, meet legal obligations, resolve disputes, enforce agreements, and maintain security logs. We also retain it for legitimate business operations. You can request deletion at any time by contacting us as described in Section 10. We may retain anonymized aggregate statistics indefinitely.

30-day availability and compliance storage. Prompt SecOps does not undertake responsibility for storing, retaining, or making scan data, scan results, or web chat image artifacts available beyond 30 days after creation, except as required by applicable law or a separate written agreement. This is not a guarantee that every record will be deleted on day 30; the data-processing retention schedule below still applies. If you need records for longer than 30 days for compliance, you must configure a SIEM connection in Settings → SIEM Integration before the applicable availability period ends. You are responsible for verifying successful delivery and configuring retention and access controls in your SIEM. SIEM delivery does not extend Necio storage and may not include every field or image artifact; separately export and preserve any required records or artifacts not delivered by the integration. Uploaded prompt attachments remain subject to their shorter 48-hour deletion period.

Data-processing retention schedule (not a storage availability commitment). Files you attach to test prompts are deleted 48 hours after upload. Raw scan and Maestro campaign prompts, model responses, private screenshot evidence, attachment delivery metadata, detailed configuration, routine activity events, and SIEM delivery events are retained for 90 days. Security-event source IP addresses are reduced to an anonymized network prefix after 90 days. Security and audit events, scan and campaign findings, and detailed result history are retained for up to 12 months, after which records are reduced to non-sensitive aggregate tombstones. Records connected to an active security incident or documented legal hold may be retained longer.

Tamper-evident security archive. To detect and investigate security incidents and to meet audit obligations, we copy security and audit records to a write-once archive: security events, records of changes to account permissions, billing and system settings, and request logs from our hosting provider. Before a record is archived we reduce IP addresses to a network prefix and remove or mask browser identifiers, URL query parameters and email addresses; archived records may still contain your account's pseudonymous user ID. Archived records cannot be altered or deleted, including in response to a deletion request, until they expire 12 months after they are archived, when they are deleted automatically. Access to the archive is limited to authorized administrators and is itself logged.

Records retained after deletion. If you delete your account or request deletion, we may retain a limited set of records where we have a legal obligation or legitimate interest to do so, including: records evidencing your acceptance of our legal terms and security-testing authorizations (consent evidence), billing and usage-metering records, tax and accounting records, and security and audit logs. These records are kept only as long as necessary for the purpose that requires them and are protected in accordance with this Policy.

5. Security

We implement industry-standard security controls including encryption at rest and in transit (TLS), hashed password storage, optional multi-factor authentication, role-based access controls, session idle timeouts, security event logging, and automated vulnerability scanning of our own infrastructure using the Necio platform. In the event of a breach affecting your data, we will notify you as required by applicable law.

6. International Processing

Prompt SecOps is based in the United States. If you access the Service from outside the United States, your information may be processed in the United States and other jurisdictions where we or our service providers operate. These jurisdictions may have data protection laws that differ from those in your location.

The model inference providers and browser automation service described in Section 3 may process scan content outside the United States. If you select a target in another jurisdiction, your prompts are sent there too.

7. Your Rights

Depending on your jurisdiction, you may have the right to:

• Access or receive a copy of the personal data we hold about you
• Correct inaccurate personal data
• Request deletion of your personal data ("right to be forgotten")
• Restrict or object to certain processing
• Data portability (receive your scan data in a machine-readable format)
• Opt out of any sale or sharing of personal data, where applicable
• Appeal a privacy-rights decision, where applicable
• Lodge a complaint with your local data protection supervisory authority, where applicable

We do not sell personal data or share it for cross-context behavioral advertising. We do not use sensitive personal data to infer characteristics about you. Because we do not sell or share personal data, universal opt-out signals such as Global Privacy Control do not change how we process your data.

Do Not Track. Some browsers send a "Do Not Track" (DNT) signal. There is no uniform industry standard for interpreting DNT, so the Service does not change its behavior in response to it. This does not affect our practices described above: we do not use third-party advertising cookies and do not track you across other websites or online services.

To exercise any of these rights, contact us at [email protected]. We may need to verify your identity before acting on a request, and we will respond within the timeframes required by applicable law. We will not discriminate against you for exercising any privacy right.

8. Age Restrictions

Necio is intended for adults and is not directed to children. You must be at least 18 years old to use the Service. We do not knowingly collect personal data from anyone under 18. If you believe we have inadvertently collected such data, please contact us immediately.

9. Changes to This Policy

We may update this Privacy Policy from time to time. Material changes will be communicated via email or an in-app notice at least 30 days before taking effect. Continued use of the Service after the effective date constitutes acceptance of the updated policy.

10. Contact Us

For privacy questions, requests, or concerns: