CITATION.cff

cff-version: 1.2.0
title: "vstack: Organizational behavior, practiced on AI agents."
message: "If you reference vstack in academic work, please cite it as below."
type: software
authors:
  - given-names: Ilhan
    family-names: Valani
    email: valani@bu.edu
    affiliation: Boston University
    orcid: "https://orcid.org/0009-0000-0000-0000"
repository-code: "https://github.com/valani9/vstack"
url: "https://valani9.github.io/vstack/"
license: MIT
keywords:
  - artificial-intelligence
  - ai-agents
  - multi-agent-systems
  - agent-debugging
  - agent-observability
  - organizational-behavior
  - after-action-review
  - psychological-safety
  - model-context-protocol
abstract: >-
  vstack is a Python library of 34 diagnostic patterns for AI agents and
  multi-agent systems, anchored in named organizational-behavior literature.
  Each pattern ports a canonical OB framework (Wharton's After-Action
  Review, Lencioni's Five Dysfunctions, Edmondson's Psychological Safety,
  Lewin's B = f(I, E), Schein's culture iceberg, and 29 others) into a
  runnable analyzer that operates on structured agent traces. The library
  exposes 13 invocation surfaces — Python imports, 59 command-line tools,
  an MCP server, a FastAPI REST service, a multi-arch Docker image,
  Claude Code skills, native adapters for ten agent frameworks (LangChain,
  LangGraph, CrewAI, AutoGen, LlamaIndex, Pydantic AI, smolagents, Agno,
  Google ADK, AWS Strands), a first-run smoke demo, and a GitHub Action —
  so the same patterns can be invoked from any agent framework or AI client
  without rewriting them.
version: "0.47.0"
date-released: "2026-06-23"
preferred-citation:
  type: software
  title: "vstack: Organizational behavior, practiced on AI agents."
  authors:
    - given-names: Ilhan
      family-names: Valani
      email: valani@bu.edu
      affiliation: Boston University
  year: 2026
  url: "https://github.com/valani9/vstack"
  version: "0.47.0"