Show HN: Policy gateway for AI agent tool calls, with a tamper-evident log

Category: security

Tags: ai-agents, policy-enforcement, audit-log, security, opa, spiffe, immudb, fail-closed

Score: 8.3/10 (Innovation: 8, Technical: 9, Documentation: 9, Utility: 7)

AIL is a policy enforcement gateway for AI agent tool calls that combines SPIFFE/SPIRE workload identity, OPA policy evaluation, and an ImmuDB tamper-evident ledger into a fail-closed pipeline. Its standout feature is radical transparency: the README explicitly lists what the system does not claim, including that prompt injection is only structurally constrained, not prevented. The project targets the emerging need for auditable, compliant AI agent deployments.

Target audience: security engineers, platform engineers, AI infrastructure teams

Repository: https://github.com/banji-007/compliance-ail · Python · MIT · 3 stars

View on Hacker News