Skip to main content

AI Safety Guardrails Policy

"

AI action safety

The lab treats AI-generated output as untrusted input until deterministic controls and, where appropriate, human approval establish that an action is permitted.

Core design principles

  • Minimize the tools exposed to an AI system.
  • Use least-privilege credentials.
  • Separate read, draft and write capabilities.
  • Validate action parameters outside the model.
  • Require explicit approval for high-impact or irreversible actions.
  • Keep a reliable audit trail of the final action and approval.

Human approval is not a decorative step

Approval should occur after the target, parameters and expected consequence are known.

The approval should apply to one specific staged action, not to a vague permission that can later be reused for a different operation.

Timeout or missing approval should not silently become approval for high-risk actions.

Prompt injection and untrusted content

Content from email, web pages, documents, support tickets or other external sources can contain instructions that should not override the system's trusted policy.

The safer architecture keeps permissions narrow and performs authorization and validation in downstream systems rather than relying on the model to judge whether an action is allowed.

AI action safety

The lab treats AI-generated output as untrusted input until deterministic controls and, where appropriate, human approval establish that an action is permitted.

Core design principles

  • Minimize the tools exposed to an AI system.
  • Use least-privilege credentials.
  • Separate read, draft and write capabilities.
  • Validate action parameters outside the model.
  • Require explicit approval for high-impact or irreversible actions.
  • Keep a reliable audit trail of the final action and approval.

Human approval is not a decorative step

Approval should occur after the target, parameters and expected consequence are known.

The approval should apply to one specific staged action, not to a vague permission that can later be reused for a different operation.

Timeout or missing approval should not silently become approval for high-risk actions.

Prompt injection and untrusted content

Content from email, web pages, documents, support tickets or other external sources can contain instructions that should not override the system's trusted policy.

The safer architecture keeps permissions narrow and performs authorization and validation in downstream systems rather than relying on the model to judge whether an action is allowed.

What this site will not recommend

  • Disabling security controls to make an integration work.
  • Using broad administrator credentials when a narrow scope is sufficient.
  • Allowing model output to directly execute destructive actions without validation.
  • Publishing real credentials, API keys or confidential production data in examples.

What this site will not recommend

  • Disabling security controls to make an integration work.
  • Using broad administrator credentials when a narrow scope is sufficient.
  • Allowing model output to directly execute destructive actions without validation.
  • Publishing real credentials, API keys or confidential production data in examples.

Useful next pages

"

Comments

Popular posts from this blog

Self-Hosted n8n Keeps Crashing: What Evidence to Collect

" Operations and recovery A self-hosted workflow platform adds infrastructure failure modes to workflow failure modes. Changing container settings, database state and workflow logic at the same time makes diagnosis harder. Resolution rule Change one boundary at a time. Preserve the failing evidence, apply the smallest safe correction, then replay the known case and check the business effect as well as the technical run status. Contents What the symptom actually proves Evidence to collect before changing the workflow Likely failure paths Resolution sequence Prevention design Verify the fix Decision table Sources and scope What the symptom actually proves A self-hosted workflow platform adds infrastructure failure modes to workflow failure modes. Changing container settings, database state and workflow logic at the same time makes diagnosis harder. A useful diagnosis begins by separating what is directly observed from what is only suspected. Execution status, HTTP co...

Contact Feedback

" Contact and feedback Useful technical feedback includes enough evidence to locate a problem while protecting credentials, customer data and confidential workflow information. Contents What to include in technical feedback What not to send How to use available site feedback channels What to include in technical feedback The URL of the page you are commenting on. The specific statement, step or link that appears wrong or outdated. The current platform and version where relevant. A link to current primary documentation if you have one. A sanitized example that does not contain credentials or personal data. What not to send API keys, passwords, access tokens or cookies. Customer names, email addresses, financial data or confidential records. Production database exports. Private webhook URLs or credential-bearing screenshots. How to use available site feedback channels Use the feedback or contact channel made available on this Blogger site when enabled. If article comme...