Evidra FAQs

Question 1

Can Evidra be used in CI/CD pipelines as well as with AI agents?

Accepted Answer

Yes, Evidra's core validation engine works seamlessly across both AI-driven workflows and traditional CI/CD pipelines. This allows you to enforce the same consistent policies and generate an evidence log, regardless of the automation source, before any execution.

Question 2

How does Evidra provide an audit trail for AI actions?

Accepted Answer

Every decision made by Evidra (allow or deny) is appended to a local, immutable, SHA-256 hash-chained evidence log. This append-only record ensures tamper detection and provides a clear, verifiable audit trail for compliance and post-incident analysis.

Question 3

How does Evidra protect my infrastructure from AI agents?

Accepted Answer

Evidra evaluates structured tool invocations (like `kubectl`, `terraform`, `helm`) against Open Policy Agent (OPA) policy bundles. It blocks destructive operations by default if they are unknown, incomplete, or target ambiguous scopes, ensuring deterministic results.

Question 4

What is Evidra?

Accepted Answer

Evidra is a fail-closed kill-switch designed for AI agents. It validates their infrastructure operations against explicit policies before execution, preventing unsafe, incomplete, or destructive actions on your systems.

Question 5

What types of misconfigurations can Evidra prevent?

Accepted Answer

Evidra ships with a curated 'ops layer' of rules that catch catastrophic misconfigurations, such as privileged containers, public S3 buckets, or open security groups. This policy layer is also extensible, allowing you to add your own custom guardrails.

Evidra

Evidra

Key Features

Use Cases

Key Features

Use Cases