Trust

Is the receipt format open-source?

Yes. github.com/genzagents/work-receipt-spec. Dual-licensed: Apache-2.0 for code (reference implementations, test vectors), CC BY-SA 4.0 for spec text. Pending Linux Foundation AAIF working group submission.

What's in the spec repo

SPEC.md — the full v0.1 spec including §3.7 Pacts, §3.8 Lifecycle, §3.4 Runtime fields. CONFORMANCE.md — 7 sections of test vectors. CONTRIBUTING.md — change process. receipt.schema.json — JSON Schema. LICENSE — split Apache-2.0 + CC BY-SA 4.0. Reference implementations in Python and TypeScript.

Why dual-licensed

Code wants permissive reuse — anyone can integrate. Apache-2.0 is the standard there. Spec text wants share-alike — any derivative format also stays open. CC BY-SA 4.0 is the standard for prose. Same pattern as W3C, OpenAPI, CNCF spec-plus-tooling repos.

Why we open-sourced it

Three reasons. (a) Defensive — competitors can't patent the format (USPTO 99462472 attempt by MerchantGuard is prior-arted by our public spec). (b) Adoption — agent marketplaces, framework vendors, complementary products can implement without depending on us. (c) Standards path — open-sourced specs are the only ones that converge to industry standards.

Linux Foundation submission

We've drafted a submission to the Linux Foundation Agentic AI Foundation (AAIF) working group proposing work-receipt-spec as a candidate format. Submission pending; if accepted, the format moves to LF governance with us as one of the contributors rather than the owner.

Adoption beyond GenZAgents

A few framework vendors have implemented the format independently for their own audit needs. The reference implementations made it easy. As the format matures, we expect more independent implementations — which is the point of open-sourcing it.

Contributing

GitHub PRs welcome. Material additions go through the CONTRIBUTING.md change process. Co-author credits to contributors. The format is small enough that contributions are manageable; large enough that ecosystem buy-in matters.

Related

Get the trust layer for your AI work

GenZAgents is the verified work-history layer above every AI provider your team uses. Sign cryptographic receipts, hand off conversations across Claude / ChatGPT / Cursor / Gemini, keep institutional AI knowledge when employees leave.

Last reviewed · 2 min read· Open spec· Changelog