SEO Office runs entirely on your own machine. API keys live in .env.local (gitignored) and never leave your laptop. There is no SEO Office server, no hosted database, and no telemetry endpoint we operate.

That said, vulnerabilities in this code can still expose your local data, your keys, or your client work. We take reports seriously.

• Anything that allows code in this repo (or in a dependency we pin) to read or exfiltrate files outside .seo-office/<active-client>/ without explicit user action.
• Anything that causes .env.local contents to be logged, written to a file outside .env.local, or transmitted to a network endpoint we did not authorize.
• Path-traversal or SQL-injection paths into the brain SQLite index.
• Prompt-injection or tool-call-injection paths in the orchestrator that let a specialist's output run shell commands the user did not approve.
• Supply-chain risks in vendored/claude-seo/ Python scripts or in the npm dependency tree that affect installation safety.

If you are unsure whether something qualifies, file it as a security report anyway. False positives are cheap; missed reports are not.

Please do not file a public GitHub issue with reproducible exploit details. Instead, use one of these private channels:

1. GitHub Security Advisories. Open a private advisory at https://github.com/AgriciDaniel/seo-os/security/advisories/new. This is the preferred path because it gives us a single thread to track the report, propose a fix, and credit you on disclosure.
2. AI Marketing Hub Pro Skool. DM the maintainer (@AgriciDaniel) in the Pro community with the keyword [security] in the subject line. We will move the conversation to a private channel within 24 hours.

Include the following so we can reproduce and assess scope:

• Affected file paths, commit hash (or version tag), and OS.
• Step-by-step reproduction (commands, inputs, expected vs actual).
• Suspected impact and any proof-of-concept artifact.

We will acknowledge the report within 72 hours, share a triage verdict within 7 days, and aim to land a fix within 30 days for high-severity issues. Status updates run on the same channel you reported through.

We do not currently backport fixes to pre-0.1 checkpoints. Update to the latest tag on main.

• Do not publish exploit details before a fix has shipped to main.
• Do not run the exploit against systems you do not own or have explicit permission to test (this includes the maintainer's own servers, demo deployments, and other members' machines).
• Do not exfiltrate other members' data while investigating.

A responsibly disclosed issue gets a credit line in the relevant release notes and, where appropriate, a thank-you post in the Pro community channel. Thanks for keeping the community safe.