2 projects
spyv
Prompt-security testing for AI engineers. Discovers every prompt across CrewAI, LangChain & OpenAI, then audits, red-teams, and (at runtime) guards them — with deterministic checks for secrets, PII & prompt-leaks that don't rely on an LLM. Works with any model.
personakit
A declarative framework for building role-based LLM agents — code reviewers, compliance officers, clinical triage, support triage. Not a personality classifier; an agent builder. Persona, frameworks, probes, red flags, and themes as data; you get a typed, cited, safety-aware agent.
