GoKawiilConduct launches open-source runtime governance layer for LLM and MCP tool calls
Conduct has released an open-source framework called Guard and a companion LLM proxy called Router that enforce a single policy across every AI agent action, from LLM calls to shell and MCP tool use. The system blocks, warns, audits, or injects rules before an action executes, using signed configuration packs and a SHA-256 hash-chained audit log to prevent tampering. A free 'Discovery' mode offers 14 days of read-only visibility into AI agent behavior before teams commit to enforcement.
GoKawiil's interpretation of the reporting above, not reported fact.
As companies give AI agents more autonomy over shell commands and tool calls, existing safety tools mostly log what happened after the fact rather than stopping risky actions in advance. Conduct's approach—verifying signed policies and chaining audit entries cryptographically—targets compliance and security teams that need provable, tamper-evident control rather than best-effort monitoring, addressing a gap as agentic AI moves into production environments.
- Conduct offers two components: Guard (policy engine) and Router (LLM proxy) working from one shared policy.
- The system fails closed by default and verifies signed configs before enforcing, unlike fail-open detection tools.
- A free 14-day read-only Discovery mode lets teams observe AI agent actions before writing enforcement rules.
Source: github.com, 2026-08-28
Published there as: “Show HN: Conduct, open-source guardrails for LLM and MCP tool calls”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.