Directory Changelog

Added to radar (AI Governance & Risk Management). Agent-run SOC 2/HIPAA readiness with infrastructure-grounded policy generation, overpromise detection, and cryptographically signed evidence packs; flat $5,988/yr per framework. Vendor claims not yet independently verified.

Added to radar (AI Governance & Risk Management). AI usage governance and policy enforcement for SMBs; EU AI Act/ISO 42001 alignment. DocsOrb UG, Germany.

Added to radar (Enterprise Security & Guardrails). In-flight interceptor for AI coding agents — prompts, shell commands, git diffs. Kolerr Lab, Inc.

Added to radar (Data Privacy & PII). On-device Mac PII/PHI redaction before content enters AI tools; open-source engine, disclosed UK operator.

Added Hathr.AI to the radar (Data Privacy & PII Redaction): HIPAA-focused AI assistant on AWS GovCloud, vendor-claimed FedRAMP High hosting and BAA. Submitted via the /submit form; pending editorial verification.

Crawl-trust fixes: /directory and /compare no longer self-redirect (307 removed); /sitemaps/jobs.xml created and registered; markdown mirrors at /tools/{slug}.md restored (served via middleware); mixed-case URLs now 301 to lowercase.

Blog cleanup: unpublished and archived 7 off-topic Contentful starter-template demo posts and 1 empty draft that predated launch. The blog now contains only editorial AI-compliance content — 16 articles.

Added /terms (Terms of Service) and /contact (business-email-only contact form). Terms cover editorial independence, AI-assisted content disclosure, no-professional-advice, liability limits, and the corrections/takedown process.

Jobs board upgraded to a full ingestion pipeline: daily automated collection from public ATS APIs (Greenhouse, Lever, Ashby) with relevance filtering, deduplication, and a public run log on /jobs for transparency.

Launched /jobs — an independent daily aggregation of AI compliance, governance, and security roles. Listings link to employers' own sites; AI Compliance Index is not affiliated with any employer listed.

Published: "An AI Model Just Hacked Hugging Face to Cheat a Test. Here Is What Defenders Do Now" — the Jul 21 agentic sandbox-escape incident and the defensive stack, mapped to the directory.

Published: "Hospitals Must Re-ID Every Off-Campus Department by 2028. The AI Doing That Work Is a HIPAA Project." — the CAA 2026 §6225 NPI mandate explained, with the HIPAA-compliant AI stack for the remediation.

Published: "The $300M Compliance Startup Accused of Faking Compliance: What Delve Teaches Every Buyer" — verified timeline with confirmed-vs-alleged ledger and a compliance-vendor verification checklist. Editorial note: Delve was evaluated and declined for directory listing.

Added to radar (Model Evaluation & Benchmarking). YC-backed agent simulation/evals; $28M Series A led by Norwest, Jun 2026.

Added to radar (Model Evaluation & Benchmarking). YC W23 open-source LLM evals/observability leader; part of ClickHouse since Jan 2026.

Added to radar (Enterprise Security & Guardrails). Identity security spanning human and AI-agent identities.

Added to radar (Enterprise Security & Guardrails). a16z-backed agent IAM; $38M raised, MCP/OAuth 2.1 standards contributor.

Added to radar (Enterprise Security & Guardrails). YC W24 LLM application security and AI vendor risk.

Published: "Vanta vs OneTrust for AI Governance: Two Different Machines Wearing One Label" — our first long-form head-to-head comparison, with G2/community sentiment and the Digital Omnibus timeline split.

Added to Frameworks & Checklists (verified). v2.0, 2026.

Added to Frameworks & Checklists (verified). The EU's GPAI compliance code, Jul 2025.

Added to Frameworks & Checklists (verified). 243 controls, 18 domains, Jul 2025.

Added to Frameworks & Checklists (verified). NIST's GenAI risk profile, Jul 2024.

Published 10 new answer pages for agent-asked questions, covering all seven categories — from LLM Guard migration to EU AI Act high-risk conformity.

Added to radar (Enterprise Security & Guardrails). Palo Alto Networks' AI security platform (v3.0, Mar 2026) — successor home of Protect AI and LLM Guard's technology.

Added to radar (Enterprise Security & Guardrails). Part of SentinelOne since Sep 5, 2025.

Added to radar (Enterprise Security & Guardrails).

Promoted to trending #3, replacing archived LLM Guard. Verified independent and active: agentic runtime security (Mar 2026), Cohere partnership (Jun 2026).

Published: "Nine AI Security Vendors Sold in Ten Months. Check Who Owns Yours." — the 2025–2026 consolidation tracker.

Ownership corrected: part of F5 since Sep 2025.

Added to radar (Enterprise Security & Guardrails).

Added to radar (Enterprise Security & Guardrails).

GitHub repository archived by owner on Jul 9, 2026, twelve months after Palo Alto Networks acquired Protect AI. Removed from trending; website repointed to protectai.com/llm-guard; description updated with migration guidance.