pact_

Proof-of-Negotiation for autonomous systems.
Cryptographic agreements, evidence, and blame logic for agents that can spend money.

[pact_ @pact_protocol]

@zenorocha We’d love to talk to you more about this!

[pact_ @pact_protocol]

Agent commerce needs a shared integrity layer.

ACTIS (https://t.co/M32Z8B2VGo) defines the standard for verifiable transaction evidence.

pact_ builds on it to generate transcripts and evidence bundles for agent transactions. Standards grow ecosystems.

[pact_ @pact_protocol]

ACTIS defines how autonomous transactions produce verifiable evidence.

pact_ is built on top of ACTIS, generating ACTIS-compatible transcripts and evidence bundles for agent transactions.

Standards create ecosystems.
pact_ helps developers use them.

[pact_ @pact_protocol]

pact_ is now a layer built on https://t.co/M32Z8B2VGo which defines a deterministic verification procedure for autonomous transaction evidence bundles

[pact_ @pact_protocol]

Autonomous agents are going live on stablecoin rails.

But execution ≠ accountability.

Pact is the constitutional trust layer:
• Proof of Negotiation transcripts
• Deterministic blame logic
• Offline-verifiable auditor packs
• Actuarial risk snapshots (Prophet)

If your agent can spend money, it should run inside a pact_ boundary.

[pact_ @pact_protocol]

History is consistent. Execution layers scale transactions. Independent standards scale trust.

Visa didn’t become internet law.
SWIFT didn’t define global liability.
Blockchains didn’t solve offchain disputes.

Agent commerce will follow the same path.
pact_ is positioned for that layer

[pact_ @pact_protocol]

Agentic commerce is coming faster than most people think.

Rails are forming. Stablecoins are institutionalizing. Agents are starting to transact.

The missing primitive isn’t payment; it’s verifiable responsibility.

If autonomous systems are going to move capital, they need replayable, deterministic evidence.

That’s the layer we’re building at pact_

[a16z crypto @a16zcrypto]

https://t.co/B2dW2Fqy9q

[pact_ @pact_protocol]

This is why we made pact_

[Tom Zschach @TomZschach]

Risk is never eliminated, only moved.

A speculative scenario from a Substack moved Visa and Mastercard nearly $100B in a few hours. The market read “agents bypass interchange” and priced in the revenue loss. Nobody priced in where the risk goes.

Interchange isn’t just a fee. It funds chargebacks, fraud liability, dispute resolution, consumer protection and regulatory compliance across every jurisdiction those networks operate in. Remove the fee, you remove the funding mechanism for all of it.

So the risk didn’t disappear. It moved to the consumer, who now has no recourse when an agent makes a bad transaction on their behalf.

And to the merchant, who now bears fraud losses directly. And to the stablecoin rails, which have no equivalent infrastructure for any of this.

$100B in market cap erased because people modeled the revenue side without modeling the risk side. That’s Risk Displacement in real time.

[pact_ @pact_protocol]

If you are building stablecoin rails, agent marketplaces, autonomous API procurement, escrow layers, or agent SaaS billing, we want to make your system replayable and institution-ready. DM us

[pact_ @pact_protocol]

Agent payment rails are emerging.
Negotiation protocols are emerging.

What does not exist yet:

A neutral, protocol-agnostic evidence standard.

That is what pact_ is

[pact_ @pact_protocol]

pact_ One

[pact_ @pact_protocol]

Pact is now published as a constitutional trust layer for autonomous agent transactions.

Two command proof from a clean clone.

Branch: pact-one-standard

[pact_ @pact_protocol]

Prophet is live in pact_ v6.0.4.

Prophet turns signed negotiation transcripts into deterministic, constitutional outcome distributions.

No scoring hacks. No hidden weights. No runtime knobs.

Same transcript + same constitution = same result.

[pact_ @pact_protocol]

If your system lets agents transact, you will eventually need evidence.

We built it.

We’re looking for more early design partners to pressure test real workflows.

DM or email to work together!

[pact_ @pact_protocol]

Project Prophet begins.

Deterministic prediction built from Pact evidence. Integrity for auditors. Optimization for agents. Shadow mode to simulate decisions before they happen.

Shipping soon.

[pact_ @pact_protocol]

Nothing in Boxer modifies evidence.

Everything is derived from sealed, hash-linked, signed transcripts.

No dashboards.
No server trust.
No subjective overrides.
Clone the repo. Run the demo. Open the Viewer.
https://t.co/3MpzZbWmBB
https://t.co/LKUcFC0FLV

[pact_ @pact_protocol]

Project Boxer ships today.

Boxer is the reputation and confidence layer built on top of pact_'s sealed evidence.

Here’s exactly what we added 👇

[pact_ @pact_protocol]

Passport Integration
Reputation is linked to agent identity anchors.
• Anchor types surfaced (KYB, service account, etc.)
• Revocation-aware signals
• Deterministic recompute across packs

Reputation is evidence-bound, not account-bound.

[pact_ @pact_protocol]

Claims ZIP Flow
You can now generate a claims package that:
• Nests the original Auditor Pack
• Includes derived Boxer signals
• Preserves transcript integrity
• Exports in a deterministic structure

Auditors see proof and derived assessment.

[pact_ @pact_protocol]

Viewer Integration that shows:
• Reputation snapshot
• Confidence level
• Clear next-step recommendation
• Unified badge color semantics
• Explicit separation of verified fact vs advisory signal

All computed locally.