Privacy Policy

Last updated: July 18, 2026

This Privacy Policy explains how AIEmployees ("AIEmployees", "we", "us", or "our"), the operator of Terminal Skills — our AI-agent-skills marketplace, open-source skills library, and Agent Playground available at terminalskills.io — collects, uses, shares, and protects personal data. "You" or "Customer" means the visitor, account holder, or contributor using Terminal Skills.

AIEmployees is the business name under which an independent entrepreneur operates; it is a sole proprietorship, not an incorporated or registered company, and not a separately registered legal entity. "Terminal Skills" and "AIEmployees" are unregistered trade names, and "Terminal Skills" is an unregistered trademark. Nothing in this Policy asserts any corporate, state, or trademark registration.

This Policy applies to our website, the Terminal Skills application (including accounts, billing, and the Agent Playground), the public skills library, and our command-line and programmatic interfaces. For the terms governing your use of the service, see our Terms of Service.

1. Overview

Terminal Skills has three parts, and how much data we process depends on which you use:

• Browsing the library (skills, use cases, install commands) requires no account. We process only standard analytics and infrastructure data.
• Using the Agent Playground, an account, or billing requires you to register. This is where we process personal data such as your email, authentication credentials, the tasks and files you submit to AI agents, and payment references.
• Contributing to the open-source library happens through our public GitHub repository, where your GitHub identity and contributed content become part of the public Git history.

We aim to collect as little personal data as reasonably possible for each of these.

2. Information We Collect

We collect personal data that you provide directly, data we collect automatically as you use the service, and a limited amount of data we receive from third parties such as Stripe.

2.1 Provided by you

• Account information: your email address (stored lowercased as your account identifier) and a password, which is never stored in plaintext — only a bcrypt hash of it is persisted. Passwords must be 8–200 characters.
• Agent Playground task prompts: the free-text tasks you type to run, inspect, or forge skills. These are stored both on disk with the run and, in truncated form (up to 2,000 characters), as a durable run summary.
• Uploaded files: files you attach to a playground run or a security scan (up to 5 files, 20 MB each). These are copied into an isolated per-run workspace and, for scans, stored alongside the scan record. The playground interface states that uploaded files are auto-deleted after 7 days and that compute happens on our servers — nothing touches your own machine.
• Skill content you submit for scanning or forging: the SKILL.md text you provide, plus the findings, verdict, and summary our tooling derives from it.
• Communications: if you contact us by email at support@terminalskills.io or via our public GitHub Issues or Discussions, we receive the message and any information you choose to include. Our contact page is a set of static links (a mailto: link and GitHub links) — there is no web contact form, and we do not store contact messages in the application.

We do not collect or store full payment-card numbers, CVV, or bank details anywhere. See Section 2.3 for how payment is handled.

2.2 Collected automatically

• Account and usage timestamps: account creation time and your last login time.
• Usage and billing counters: credit balance, cost accrued today, run/scan/forge counts (daily and lifetime), total cost, and quota-reset timestamps — used to operate the credit wallet and enforce plan limits.
• Per-run diagnostics: run status, engine and model used, execution mode, which skills were routed and activated, generated artifacts, token totals, cost, corrections, and any error. A full per-run event transcript (the conversation and tool-event stream) is written to disk for each run and is viewable by our administrators for support, safety, and debugging.
• IP address (transient): we read your IP address from request headers only for rate limiting. It is held in an in-memory store on a rolling 60-second window and is not written to any database or persistent log.
• Analytics data: we load Google Tag Manager on every page (see Section 7). The tags configured inside our Tag Manager container may collect page views, browsing behavior, device/user-agent information, IP address, and set cookies.

We do not perform UTM/attribution capture or server-side logging of personal request data in our application code.

2.3 From third parties

• Stripe (payments): when you buy credits or subscribe to Pro, Stripe returns a customer ID and (for subscriptions) a subscription ID, which we store on your account. We can also fetch your current subscription status, renewal date, and cancellation state live from Stripe to show them on your account page. We keep an internal ledger of processed Stripe webhook events (event ID, type, your user ID, and credits granted) purely for idempotency. Card details are entered directly on Stripe-hosted checkout and never reach our servers.
• GitHub (contributor identity): the "author" attached to an open-source skill is a public GitHub username supplied through our public repository — it is part of the open catalog, not collected through platform login. Our homepage also fetches the public star count of the repository (no personal data).

No third-party social login exists. Authentication is email + password only — there is no Google, GitHub, or other OAuth/single-sign-on login.

3. How We Use Information

We use personal data to:

• operate, maintain, and secure the website, the skills library, and the Agent Playground;
• create and manage your account and authenticate you;
• run the AI agents you request, meter their compute cost against your credit wallet, and enforce plan limits and cost caps;
• process payments, manage subscriptions, and maintain accurate billing and usage records;
• provide support and respond to your messages, bug reports, and contributions;
• detect, prevent, and investigate abuse, fraud, and security incidents (including rate limiting and refusing skills flagged as unsafe);
• understand which skills and pages are useful, in aggregate, so we can prioritize work;
• comply with legal obligations and enforce our Terms of Service.

4. Legal Bases for Processing (GDPR)

Where the EU/UK GDPR applies, we rely on the following legal bases:

• Performance of a contract — to provide accounts, the Agent Playground, and billing you have signed up for.
• Legitimate interests — to secure the service, prevent abuse, maintain diagnostics, and improve the product, balanced against your rights.
• Consent — where we obtain it. As disclosed in Section 7, analytics tags currently load without a consent gate, and we do not yet operate a consent-management mechanism; we are working toward one, at which point consent (where legally required and captured) would serve as the basis for non-essential analytics cookies and similar technologies.
• Legal obligation — to meet tax, accounting, and other legal requirements.

5. How We Share Information — Sub-processors

We do not sell personal information for money. We share personal data with the service providers ("sub-processors") below, each of which processes data on our behalf or as an independent controller under its own privacy policy. This list reflects only the vendors actually used by Terminal Skills.

Our database (MongoDB) and authentication are self-hosted on our own infrastructure — we do not use a managed database service (such as MongoDB Atlas) or a third-party identity provider. We do not use a third-party object-storage or CDN service for your uploaded files; run workspaces are stored on the platform's local filesystem.

We may also disclose personal data to comply with law, enforce our agreements, protect the rights, property, or safety of AIEmployees, our users, or the public, or in connection with a merger, acquisition, or sale of assets.

6. International Data Transfers

Our hosting infrastructure (Hetzner via self-hosted Coolify) is provided by a Germany-based provider, while several of our sub-processors (Stripe, GitHub, Google Cloud with a "global" region) process data in the United States and other regions. Where personal data is transferred out of the EEA, UK, or Switzerland, we rely on appropriate safeguards such as Standard Contractual Clauses (and the UK Addendum / Swiss adaptations) where required. The exact data-center region of our hosting is available from us on request.

7. Cookies & Analytics

We do not currently display a cookie-consent banner, and analytics is not gated behind consent. We want to be transparent rather than overstate our controls:

• The only cookie our own application sets is ts_session — a first-party, httpOnly, SameSite=Lax authentication cookie (marked Secure in production) that holds your signed session and expires after 7 days. It is strictly necessary to keep you signed in.
• Google Tag Manager (container GTM-KZ98N24T) loads on every page for every visitor as soon as the page loads. Whatever tags are configured inside that container — which may include Google Analytics — run and may set their own first- or third-party cookies (for example _ga) and collect analytics data. The specific tags and cookies are defined in the remote container, not in our source code.
• We also store a few functional flags in your browser's sessionStorage (for skipping the intro animation, restoring scroll position, and preserving a pending playground task across login). These stay in your browser and are never sent to our servers.

We do not currently operate a consent-management mechanism or automatically process Global Privacy Control (GPC) signals; analytics tags load for all visitors as described above. Your only current control is to block or clear these cookies through your browser settings or a tracker-blocking extension, and the site remains fully functional without analytics. Our fonts are self-hosted, so browsing the site does not send your IP to Google for fonts. For a categorized breakdown of every cookie and storage key we use, see our Cookies Policy.

8. Data Retention

• Account records, security-scan records, and the Stripe billing-event ledger are retained indefinitely; we do not currently apply automatic expiry to them.
• The MongoDB run summary (task prompt excerpt, cost, tokens, skills, status) is kept as durable usage and cost history and is retained indefinitely.
• On-disk run workspaces, run files, and event transcripts are intended to be short-lived; the Agent Playground interface tells users uploaded files are deleted after 7 days. We enforce this deletion operationally rather than through automated code committed today, and are working to make the schedule automatic and precise.
• Transient data — analytics data is retained per Google's settings; the star-count cache lives ~120 seconds; the rate-limiting IP window is 60 seconds in memory; API cache windows are 60–300 seconds.

We periodically review the personal data we hold and delete or anonymize data that is no longer needed for the purposes described in this Policy, consistent with the legitimate-interests and legal-obligation bases in Section 4.

Because our current retention of accounts and run summaries is indefinite, and self-service deletion is not yet available (see Section 10), you can ask us to delete your data by emailing us; we will honor valid requests as described below.

9. Data Security

We protect personal data with the following measures, described accurately rather than aspirationally:

• Password hashing: passwords are hashed with bcrypt (cost factor 10). Plaintext passwords are never stored.
• Sessions: authentication uses stateless JSON Web Tokens (HS256, signed with a server secret of at least 32 characters), carried in an httpOnly cookie and expiring after 7 days. Because sessions are stateless, an individual token cannot be centrally revoked before it expires; logging out clears the cookie, and administrative role changes are re-checked against the database on every privileged request.
• Transport: HTTPS/TLS is terminated by our hosting reverse proxy; the session cookie is marked Secure in production.
• Access control: we use role-based access control with two roles (user and admin). Administrative endpoints re-verify the admin role against the database on every request, so a revoked administrator loses access immediately.
• Secrets: application secrets are supplied through environment variables and are not stored in the database.

We do not currently apply field-level or application-layer encryption to data at rest; any at-rest protection is provided at the database/host layer. No method of transmission or storage is completely secure, and we cannot guarantee absolute security.

10. Your Privacy Rights

GDPR (EEA / UK)

If you are in the EEA or UK, you may have the right to access, correct, delete, port, restrict, or object to the processing of your personal data, and to withdraw consent where we rely on it. You may also lodge a complaint with your local supervisory authority.

US State Privacy Rights

If you are a resident of a US state with a comprehensive privacy law (such as California, Colorado, Connecticut, Virginia, Texas, and others), you may have the right to know/access, correct, delete, and port your personal data; to opt out of the "sale" or "sharing" of personal data and of targeted advertising and certain profiling; to control sensitive data; to appeal a denied request; and not to be discriminated against for exercising these rights. We do not sell personal information for money. To the extent that loading analytics could be considered "sharing," note that we do not currently operate a consent-management mechanism or automatically honor the Global Privacy Control (GPC) browser signal; you may opt out by contacting us at support@terminalskills.io or by blocking cookies in your browser or with a tracker-blocking extension.

Do Not Sell or Share

We do not sell your personal information for money. Because we load Google Tag Manager (see Section 5), that activity could, under some US state laws, be considered "sharing" for targeted advertising. As we do not yet operate a consent-management mechanism or automatically honor GPC signals, you can exercise your Do-Not-Sell/Share opt-out by blocking cookies in your browser or with a tracker-blocking extension, or by emailing us at support@terminalskills.io (subject "Attn: Privacy"). For more detail, see our Do Not Sell or Share My Personal Information page.

How to exercise your rights

Because self-service account deletion and data export are not yet built into the product, please send rights requests to support@terminalskills.io (subject line "Attn: Privacy") or through our contact page, or to AIEmployees at aiemployees.us. We will verify your request and respond within the timeframes required by applicable law (generally one month under GDPR, 45 days under US state laws, extendable where permitted). Note that public open-source contributions are handled on a best-effort basis (see Section 13).

11. Children's Privacy

Terminal Skills is not directed to children. You must be at least 18 years old to create an account or use the Agent Playground. We do not knowingly collect personal data from anyone under 18; if you believe a minor has provided us personal data, contact us and we will delete it.

12. Third-Party Services & AI

The Agent Playground runs AI agents on third-party AI models (Google Gemini via Vertex AI). AI output is inherently probabilistic and may be incomplete, inaccurate, or unsafe. You are responsible for reviewing any code, command, skill, or other output an AI agent produces before relying on it or executing it — especially in a production environment. Avoid submitting sensitive personal data, secrets, or confidential material into task prompts or uploaded files, since that content is transmitted to our AI sub-processor to perform the task you request. Skills you obtain frequently invoke other third-party platforms and APIs; each is governed by its own terms and privacy policy.

13. Open-Source Contributions

When you contribute a skill, use case, fix, or other content to our public GitHub repository, your GitHub username, commit metadata, and the content you contribute become part of the public open-source repository and its public Git history, which is retained indefinitely and mirrored by others outside our control. Public Git history cannot be selectively rewritten without breaking project integrity, so requests to erase public contributions are addressed on a best-effort basis only. Skills authored by the Agent Playground's "forge" mode are, when contribution is explicitly enabled, published to the public repository as deliberately generalized, PII-stripped content and labeled as agent-generated.

14. Programmatic Access (CLI, MCP, and API)

The terminal-skills command-line tool and our public MCP server and API expose only the public skills catalog; they do not read your account, playground runs, or billing. Where programmatic access touches account-scoped data, it is limited to your own account and remains subject to this Policy, our Terms of Service, and our Acceptable Use Policy.

15. Changes to This Policy

We may update this Policy as the product evolves. The "Last updated" date at the top reflects the most recent revision, and material changes will be highlighted on the website. Continued use of the service after an update constitutes acceptance of the revised Policy.

16. Contact

Questions about this Policy or your personal data can be sent to support@terminalskills.io (subject "Attn: Privacy"), through our contact page, or to AIEmployees at aiemployees.us. You can also review our related policies: Terms of Service, Cookies Policy, Data Processing Addendum, Acceptable Use Policy, and Do Not Sell or Share My Personal Information.