Tag
The author built Scyvera, a runtime enforcer tool that governs AI agents by enforcing declared permissions and logging decisions to address governance gaps in production systems.
The article warns against using model-generated narration as the authoritative audit log for AI agents, advocating for persisting raw tool call data instead, and suggests a simple diff check to catch discrepancies.
A developer shares their open-source HealthClaw Guardrails project that enforces safety guardrails for LLM agents accessing real health records, including PHI redaction, audit logging, and human-in-the-loop confirmation, with a conformance endpoint for testing.
Open-sourced a pre-call spend gate and tamper-evident audit log to prevent runaway API costs from recursive loops in autonomous agents, MIT-licensed with Postgres+TypeScript demo.
Halo is an open-source tool providing tamper-evident runtime records for AI agents, enabling auditable logs of agent actions that can be verified by third parties.
The author discusses a failure mode encountered while building Sentience Governor, a Python library for Claude Code that monitors agent actions and produces audit reports. The AI sometimes reconstructed explanations from raw traces, blurring the line between measured facts and probabilistic interpretation.
git-mem is a memory solution for AI agents that leverages git and redis for performance and audit logs, including a web UI and minimal dependencies.
Discusses the importance of proper audit logging for AI agents, emphasizing the need for append-only, hash-chained logs that prevent tampering, rather than storing logs in the same writable application database.
GetMCP is a self-hostable open-source tool that brings zero-trust security to AI agents by providing per-request audit, per-agent revocation, policy enforcement, and human-in-the-loop approvals for API calls. It generates MCP servers from OpenAPI specs and acts as a streaming proxy with tamper-evident audit logs.