AI agent safety and governance turns operational boundaries into written, reviewable controls: permissions, defensive handling, audit evidence, safety cases, and readiness gates.
Who it is for: Agent operators, safety leads, engineering managers, and reviewers who need explicit controls and evidence before or after agent actions.
Start with the permission and safety-gate products to define authority. Add prompt-injection defense for agents that read untrusted content, then use audit trails, safety cases, and deployment-readiness records to review evidence.
Start with the permission and safety-gate products to define authority. Add prompt-injection defense for agents that read untrusted content, then use audit trails, safety cases, and deployment-readiness records to review evidence.
Choose a bundle when: several exact members independently fit. Start with Wave 2 Agent Trust & Safety Stack, or use Agent Tool Permissions vs Safety Gates for a guided comparison.
The product type and summary below show how each option differs. Open a product page for exact scope, tiers, buyer fit, and bundle membership.
A written permission system for agent tool use: risk tiers, deny-by-default matrices, allowlist/blocklist policies, and an audit-log schema.
A written safety system for agents acting on your behalf: permission tiers, human-approval checkpoint patterns, risk classification, do-not-touch boundaries, and rollback playbooks.
Defensive patterns against prompt injection: a content trust ladder, instruction hierarchy model, untrusted-content policies, detection heuristics, and regression test suites.
A written audit-trail standard for agent actions: an 8-class action taxonomy, per-class JSON entry schemas, append-only integrity rules (tamper-evident, honestly not tamper-proof), redaction rules, and sampling plans.
A written safety-case system for agent deployments: claim-argument-evidence framing, safety-case and evidence-registry JSON schemas, hazard analysis, and review records.
An evidence-gated go/no-go system for agent deployments: capability/safety/rollback/monitoring/ownership gate families, evidence-required checklists, and sign-off records.
A written governance system for what agents may remember: memory classification, a binding never-store list, retention schedules, and audit scorecards.
These products are document, schema, policy, and checklist systems. Your own agent runtime or workflow platform must apply the controls and produce the evidence.
They do not provide runtime enforcement, monitoring, certification, legal advice, or a guarantee that every failure or attack will be prevented.
