v0.2.0

Security-hardening release from a line-wide code review. Highlights: secure-by-default approver gate for high/critical operations (init seeds a starter rules.yaml), percent-encoded URL path segments, single-sourced version, governance persistence + CLI write-path tests. See CHANGELOG.md for the full list. BREAKING: fresh installs with no rules.yaml now deny high/critical writes until an approver is recorded.

AI-Guardian v0.1.1

Patch: config.yaml now honors the *_AIOPS_HOME env var; CLI writes are now audited + undo-recorded through the governance path (previously MCP-only). See CHANGELOG.

ai-guardian v0.1.0

Governed observability + governance for on-endpoint local LLM (Ollama) usage — 18 MCP tools. Model allow/deny policy + provenance digest-drift, a deterministic offline prompt scanner (secrets / PII with Luhn / source-code / jailbreak, weighted risk band, redacted findings), and route-through guards (guarded_generate / observe_chat) that scan + policy-gate + record before calling the model. The complement to IGEL AI Armor. Vendored governance harness; remove_model is high-risk with dry-run + double-confirm. Preview — v0.1 passive inventory + opt-in route-through (transparent proxy is v0.2).