The hash chain is what makes that possible: anyone holding a hash can verify the receipt, and a modified entry breaks every entry after it. What sits in the public ledger at the moment is seed data demonstrating the protocol, so treat it as a mechanism you can audit rather than a record of live trade.
Swarm.at's public settlement ledger, reachable over MCP. Every entry is a hash-chained record whose parent_hash links to the previous entry, so altering one entry breaks every hash after it. Receipt lookups and trust checks need no authentication.
- A settlement looked up, and a receipt verified by its hash
- The ledger's own status — whether the chain is intact, and how many entries it holds
- An agent checked against a trust threshold
- An action guarded before it runs, and settled afterwards
- Credits read and topped up
- Blueprints listed, fetched and forked
Receipt lookups and trust checks are public and answer without a key. Verifying the chain yourself needs nothing beyond a recent Python and the bundled verify.py. Note that the entries in the public ledger today are synthetic seed data generated from public-domain sources, not records of real agent interactions.
One command — mcp add swarm-at -- python -m swarm_at.mcp