01
Public access and operational logs
Public pages do not require identity and do not run AI online. Website and API infrastructure may receive IP addresses, timestamps, request paths, browser information and error logs for delivery, rate limiting and diagnostics. Account-free access does not mean no access data is generated.
02
Local projects and the registry
The local runtime manages project configuration, the cached catalog index and version locks. Search runs locally over the cached index; queries never leave your machine. The client only contacts the GitHub catalog repository to download the index and skill files. This mechanism does not require uploading an entire project, secrets or full conversations to any server. External model providers process data under the user's chosen service.
03
Inquiry purpose and retention
The inquiry form stores product, company, HTTPS website, contact name and details, request, consent and submission ID for authorized operators to assess, contact and track handling. Once completed or marked invalid, personal contact details are cleared by the cleanup process after a 180-day retention period. Business status and audit records without personal data may remain. This period does not start at submission while an inquiry is being followed up.
04
Internal access and contact
Skill content is managed in a GitHub catalog repository; access is governed by the repository and GitHub accounts. Raw evaluation records never enter the public catalog repository; only approved summaries are public. To correct catalog content, open a repository issue or use an established contact channel.