Crawler telemetry notice for jkatzy.nl

Purpose
Requests routed to the automated-agent API are recorded to study crawler and
AI-agent access patterns and to support abuse analysis. This measurement
service is observe-only and does not automatically block visitors.

Data collected
- UTC timestamp and a random event identifier
- direct source IP supplied by the trusted reverse proxy
- HTTP method and path
- bounded User-Agent and Signature-Agent header values
- the crawler-detection signal used by the reverse proxy
- the self-reported agent-name, query, and access purpose submitted to /agent-api
- validation outcome and HTTP status

The service does not collect request bodies, cookies, authorization headers,
system prompts, conversation history, or credentials. Agent names and access
purposes are self-reported and are not treated as verified.

API usage guide
The request format, parameter limits, response fields, and example are available
at https://jkatzy.nl/agent-api-guide.txt.

Retention and access
Individual event files have no age-based expiry or per-day cap. They have a
dedicated 100 GiB filesystem quota and are accessible only to the server
operator. If that quota is exhausted, collection stops rather than deleting
older events. The API container has no published port and no outbound network
access.

The reverse proxy separately records the direct source IP, source port, method,
path, status, timing, and response size for all site requests. Query strings,
headers, cookies, and bodies are omitted. These access logs have no age-based
expiry and use a shared 50 GiB system-journal capacity boundary. They survive
proxy container replacement, but finite local storage cannot guarantee
permanent archival once that boundary is reached.

Do not submit personal data, credentials, private conversation history,
system prompts, or other secrets to /agent-api. Contact the site owner through
the contact details on https://jkatzy.nl/ with questions or deletion requests.
Telemetry events can be located and deleted by exact event ID or source IP.