AgentKits

Updated June 2026: OpenAI is retiring AgentKit's visual builder, Google reframed ADK as an OTel-native execution framework, and Inngest ships durable agent workflows. Architecture, memory, pricing, and fit compared.

Paste two versions of an AI agent (agentaz.json or system prompt) and get a deterministic governance diff: which tools lost approval gates, whether the Trust Level escalated, and what changed - with evidence, the same verdict every run. Free, no API key.

How AgentAz design-time governance dimensions map to NIST AI RMF 1.0, ISO/IEC 42001:2023, and the OWASP Top 10 for Agentic Applications (ASI01-ASI10). A free reference for compliance questionnaires - with honest scope gaps marked.

Pulls logs, metrics, traces, and recent changes into one timeline, ranks root-cause hypotheses, and gates any published RCA behind a human. Diagnoses, never remediates.

Maps each change to your control catalog, assembles tamper-evident evidence, and gates filing behind a human - it records, it doesn't decide compliance.

KitForge by AgentKits generates LangGraph agent scaffolds whose guardrails actually enforce: authority budgets that block, an HMAC-chained audit log that fails on tamper, and human-in-the-loop gates that halt. Free, Python, MIT.

Check any AI agent before it ships - including agents generated by ChatGPT, Claude, or Cursor. Paste a system prompt or agentaz.json for a deterministic governance score: pass/fail gates, failure scenarios, and a concrete fix block. Free, no login, prompt

An IT helpdesk agent that resolves common tickets from your runbooks, verifies identity for sensitive actions, and escalates security incidents and privileged access.

An invoice-extraction agent that parses invoices into validated JSON with per-field confidence, validates totals, and flags illegible or mismatched fields for review.

An interview-summary agent that turns notes into evidence-based feedback against role criteria, screens out biased factors, and never makes the hiring decision.