AI agents: start at /agents/; use ontology.json, relationships.yaml, llms.txt, and llms-full.txt.
Portfolio by problem

Solutions

Start with the four connected work areas, or browse individual projects below.

Explore the need each part of PAICE.work PBC serves. These are tools, references, protocols and frameworks at different stages, not a claim that every capability is a validated product. Talk with Sam about supporting the work.

Understand and improve AI collaboration

Measurement of human-side collaboration skill and the agentic web, with a supporting framework for understanding aggregated intelligence.

PAICE.work

A snapshot of human-AI collaboration skill on the human end.

Who it's for: People learning, teaching or assessing how they work with AI.

Visit paice.work →

Siteline

Making the agentic web scoreable, one website at a time.

Who it's for: People responsible for websites that AI agents use on someone's behalf.

Visit siteline.to →

AI Posture

Aggregated Intelligence Posture: a supporting framework for understanding and communicating aggregated intelligence. Practical applications remain to be developed.

Who it's for: People seeking a mental model for understanding aggregated intelligence.

Visit aiposture.org →

Connect legal information with source evidence

A connected, machine-readable graph of AI laws, obligations, incidents, and the shared schema underneath them.

Every AI Law

Structured references to AI-related obligations and their sources, connected through the Legal Graph.

Who it's for: People researching AI obligations and the evidence behind them.

Visit everyailaw.com →

PubLedge

Open recordkeeping conventions for fact-specific interpretations and administrative issuance records.

Who it's for: People who issue or rely on civic and regulatory publication records.

Visit publedge.org →

AI Incident Law

Public-record legal and regulatory matters involving AI, with allegations and outcomes kept distinct.

Who it's for: People investigating AI-related matters and their legal evidence.

Visit aiincidentlaw.org →

Obligation First

The shared vocabulary and validation contract connecting the Legal Graph while preserving each source's meaning.

Who it's for: People building interoperable legal-information systems.

Visit obligationfirst.org →

Build agents and services others can trust

Open standards and templates for the agent layer: operational limits, provenance, instruction integrity, peer coordination, and structured knowledge.

Graceful Boundaries

Discoverable limits, understandable refusals and useful next steps for humans and agents. An enabling standard for Siteline.

Who it's for: Services and agent builders that need to communicate operational boundaries.

Visit gracefulboundaries.dev →

Skill Provenance

Portable identity, integrity and version history that help make work with agent skills governable.

Who it's for: People authoring, distributing or using skills across different systems.

Visit skillprovenance.dev →

Turnfile

Session-scale coordination with recorded positions, preserved disagreement and human resolution.

Who it's for: People coordinating consequential work across agents and sessions.

Visit turnfile.work →

AIDR

One portable decision record for the question, positions, objections and human decision.

Who it's for: People who need to inspect how an AI-assisted decision was reached.

Visit aidr.work →

GuideCheck

Human-reviewable installation instructions, designed to make delegated software installation inspectable. Conformance does not guarantee safety.

Who it's for: People asking agents to install software, and the publishers supplying instructions.

Visit guidecheck.org →