After the agreement: tracing commitment breaches across organisations (a governance view)

[AmitDeb]

Hi all. I've been following #1725 (Concordia), #1737 (A2CN) and #2124 (compensation semantics). Together they cover negotiation, mandate verification, signed agreement records and undo. I'd like to raise a question about what comes after a commitment is signed, from the side of the organisations that have to govern and audit these systems.

The scenario

A buyer's procurement agent agrees delivery terms with a supplier's agent. The supplier then depends on a carrier's agent to meet them. The carrier misses its date, the supplier delivers late, and the buyer's process downstream fails. Each organisation's agent did what it was asked, and each has its own signed records. What's still hard:

1. Breach attribution along a chain. Which commitment failed first, and under whose mandate? A receipt shows that an agreement was made. It doesn't show that the promise was later broken, or that one promise depended on another.
2. Live status of an obligation. Between "agreed" and "settled", is a commitment active, at risk or violated? Today this sits in each party's internal systems, not in anything both sides can see.
3. Evidence for the governance owner. An auditor or risk owner wants to check the system: which commitments each workflow can make, under which limits, and what happened when they failed. They don't want to replay every message. This is how ISO 27001 governs information security, and I think connected agentic workflows need the same approach.

What I'm proposing to explore

A small post-agreement lifecycle for a commitment (conditional → active → satisfied | violated, plus expired), with a link from a child commitment to its parent so a breach can be traced across organisations. I'd like this to sit on top of what already exists:

• the agreement record and references[] from Concordia (a depends_on or similar relationship for nested commitments?)
• the mandate object from A2CN for authority and limits
• the compensation declarations from Transactional semantics gap: canceled is not compensated: reference schema for a commitment layer #2124 for remedies

I call this framing the Workflow Commitment Protocol (WCP), but I'm not proposing another wire format if these can carry it.

Questions

1. Is post-agreement status (active, at risk, violated) already planned in Concordia or A2CN, or is it out of scope for them?
2. Would a parent/child dependency between commitments fit as a references[] relationship, or does it need its own artifact?
3. Is there appetite for a governance and audit profile, meaning the evidence an organisation's risk owner would need, layered over these primitives?

Background: The Missing Layer: A Workflow Commitment Protocol for Agentic Organisations · preprint: doi.org/10.5281/zenodo.23740006

Happy to be pointed to prior work I've missed.

I drafted this with AI assistance. The ideas and questions are mine.