For AI agents: llms.txt is a short guide that sends you to the rules your problem needs, and rules.json holds every rule of this playbook as JSON.

How to do it

An attacker who gets into a server deletes its local logs first, and a lost disk takes them too. Run a shipper such as Fluent Bit, Vector or the OpenTelemetry Collector, or let the platform collect standard output. Send the lines over TLS to a store that the server may only add to. Let only named people read the store, and record each time they do. Audit events need a stronger store, as Make the audit log impossible to change says. The store keeps lines for the period in Keep logs for a fixed time.

Examples

Using these examples: they show one way to follow this rule. Keep the behaviour and the check, not the syntax. Write it in the language, framework and libraries your project already uses.

YAML example

# OpenTelemetry Collector on the server: read the log files and send them out over TLS
receivers:
  filelog:
    include: [/var/log/shop/*.log]
exporters:
  otlphttp:
    endpoint: https://logs.example.com
service:
  pipelines:
    logs:
      receivers: [filelog]
      exporters: [otlphttp]

INI example

# systemd-journal-upload: send the journal to a collector on another machine
[Upload]
URL=https://logs.example.com:19532
ServerKeyFile=/etc/ssl/private/journal-upload.pem
ServerCertificateFile=/etc/ssl/certs/journal-upload.pem
TrustedCertificateFile=/etc/ssl/ca/logs-ca.pem

How to check

Write a test line on the server. It appears in the central store within 60 seconds. A delete request for that line, sent with the server's own credentials, is refused.

Sources

• OWASP Logging Cheat Sheet

Related rules

• Log what every agent session does in AI development
• Make the audit log impossible to change in Compliance
• Keep logs for a fixed time in Operations