Execution kernel
gove-zone
Python 3.11+ · Beta · release candidate
Require a valid decision receipt before an integrated tool call can run.
DEVELOPER RESOURCES
Your first
governed action.
Choose a workflow, follow its local setup, and inspect the result. Public repository references below are pinned to the reviewed commits; browse 13 framework adapters and workspace architecture for the full 18-repository catalog.
START WITH ONE WORKFLOW
Choose an evaluation.
Each guide identifies the setup, controls to check, and evidence to retain. These are local evaluation paths, not a hosted service.
gove-zone
Control a tool call
Require authorization before an integrated action runs.
Start with one tool and an explicit execution boundary.Open evaluation guide Beta · release candidateACGS-LiteCheck a Python action
Evaluate a request against your configured policy.
Start with a local policy check, then inspect governed wrappers.Open evaluation guide BetaCFT Governance PackReview a cloud plan
Check a Google Cloud Terraform plan before apply.
Start with plan JSON; your pipeline enforces the result.Open evaluation guide Early source companionAlready know your component? Go to the source guides
Python policy library
ACGS-Lite
Python 3.10+ · Beta
Start with configurable policy checks and governed Python wrappers.
Multi-agent coordination
ACGS Swarm
Python 3.11+ · Mixed maturity · optional research
Coordinate task ownership, peer validation, and governance evidence.
Self-hosted HTTP service
Enhanced Agent Bus
Python 3.11+ · Evaluation · operational setup required
Operate governed messaging and workflow endpoints as a separate service.
Infrastructure policy companion
CFT Governance Pack
Python 3.10+ · Early source companion
Check Google Cloud Terraform plan JSON and retain control evidence.
SHARED REFERENCES
Understand the boundary.
Claim & evidence ledger
Dated source, release, CI, and limitations index for the product family.
Review evidenceFramework adapters (13)
PyPI packages for FastMCP, LangChain, OpenAI Agents, Claude Agent SDK, CrewAI, Pydantic AI, and more.
Browse adaptersArchitecture & bounded contexts
Receipts, grants, SQLite EffectStore lifecycle, TLA+ models, and persistence boundaries across 18 repos.
Explore architectureGOVERNANCE FIELD GUIDE
Find the question. Define the boundary.
Governance Q&A
Search authored answers with direct source references.
ExploreFailure modes
Find warning signs and controls across 19 workflow patterns.
ExploreGovernance patterns
Choose an operating mode and supporting controls.
ExploreFor agents
Copy a framework and download a blank task brief.
ExploreTrust center
Review security, privacy, and evidence in context.
ExploreWhy ACGS exists
Read the idea behind the project.
ExploreEVALUATIONS AND WORKFLOW REFERENCES
Explore the next layer.
Additional source packages, local examples, and design references. Each page states what is implemented and what remains open.
Governance Evaluation
Source MVP for authority, policy, and evaluation evidence.
Read the referenceClinicalGuard
Healthcare assessment prototype for source evaluation.
Read the referenceLegal workflow example
Mock drafting, matter access, and review boundaries.
Read the referenceHermes integration
Source middleware around the tool-call lifecycle.
Read the referenceEU AI Act mapping
Engineering evidence mapping with explicit gaps.
Read the referenceCredential-governance concept
A design reference; implementation remains unverified.
Read the reference