chore(orchestration): create scaffold for swap anything asyncFlow #11304

anilhelvaci · 2025-04-21T06:24:23Z

Description

Crosschain Swaps (XCS) is a CosmWasm contract on Osmosis that lets users swap tokens over incoming IBC transactions and send them to their destination. This destination could be on Osmosis, the incoming chain, or any other chain. What enables XCS contract to handle this complex operation is two other CosmWasm contracts, namely, SwapRouter and Crosschain Registry.

SwapRouter: In charge of keeping track of routes between tokens. Especially useful when there's no direct pool between the assets the user wants to swap.
Crosschain Registry: Responsible for keeping the information related to figuring out what chain a certain address belongs to, and what is the IBC channel between those chains

swap-anything.contract.js is an orchestration that's very similar to send-anywhere.contract.js. The contract logic just makes sure it sends a valid ICS-20 message with a special focus on;

The destination chain is ALWAYS Osmosis.
The memo is a special JSON string that can be picked up by IBC hooks on Osmosis so it can be forwarded to the XCS contract.

Here's the helper function we use to build that memo:

agoric-sdk/packages/orchestration/src/examples/swap-anything.flows.js

Lines 42 to 61 in 86ba674

    
           const buildXCSMemo = swapInfo => { 
        
             const memo = { 
        
               wasm: { 
        
                 contract: swapInfo.destAddr, 
        
                 msg: { 
        
                   osmosis_swap: { 
        
                     output_denom: swapInfo.outDenom, 
        
                     slippage: { 
        
                       twap: { 
        
                         window_seconds: swapInfo.slippage.windowSeconds, 
        
                         slippage_percentage: swapInfo.slippage.slippagePercentage, 
        
                       }, 
        
                     }, 
        
                     receiver: swapInfo.receiverAddr, 
        
                     on_failed_delivery: swapInfo.onFailedDelivery, 
        
                     nextMemo: swapInfo.nextMemo, 
        
                   }, 
        
                 }, 
        
               }, 
        
             };

Security Considerations

Currently we use one localchainAccount to handle incoming offers. I'm not sure this could be limiting factor or not. Just find it beneficial to point out.

Documentation Considerations

If we decide to use this contract in a product, it could be a good idea to document the XCS contract's behavior over certain conditions in order to help developers working on swap-anything.contract.js.

Testing Considerations

Here are the test cases covered at the new job included in the multichain-e2e CI workflow;

Swap BLD against OSMO, receiver on Agoric
Swap BLD against OSMO, receiver on CosmosHub
Swap OSMO against BLD, receiver on Agoric
Swap output sent to an unknown address, swap should fail, user purse recovered
Verify osmosis xcs state

A word on the test environment
We create a new pool between BLD <-> OSMO on Osmosis and configure crosschain_swaps, crosschain_registry, swaprouter contracts according to the IBC connections established by Starship and the pool we have created. In the next phase, we consider adding new pools and exploring test cases for PFM.

Reasoning for a dedicated starship config.yaml (config.xcs-swap-anything.yaml)
We will need to see the logs of the XCS (and other CosmWasm) contracts, if anything goes wrong. XCS code prints logs to debug level only. In order to avoid unnecessary log output for other tests, we've used our own config.

Upgrade Considerations

None

cloudflare-workers-and-pages · 2025-04-21T06:27:24Z

Deploying agoric-sdk with Cloudflare Pages

Latest commit:	`e962549`
Status:	✅ Deploy successful!
Preview URL:	https://2cc6507c.agoric-sdk.pages.dev
Branch Preview URL:	https://jorge-8863-refactor-test-too.agoric-sdk.pages.dev

View logs

closes: #11374 ## Description This PR implements the design described at #11374. The need for such a design is surfaced during when we tried to handle **any** denom incoming to a contract's address hook. We are trying to unblock that functionality with this solution. ### Security Considerations None. ### Scaling Considerations No concerns about this. ### Documentation Considerations None. ### Testing Considerations We some unit tests that showcases this method's functionality. Ideally there should also be multichain test where a contract uses this method to make sense out of some incoming IBC transfer in its address hook. We do have an on-going effort for that. See #11304 ### Upgrade Considerations None.

dckc · 2025-05-19T15:16:38Z

34 commits... some include "WIP"... @anilhelvaci it looks like you plan to squash; that is: review by commit is not the way to go. Can you confirm?

dckc · 2025-05-19T15:24:07Z

other CosmWasm contracts [on Osmosis]: Crosschain Registry ...

Interesting. How are changes to the registry governed?

We were going to put chainInfo in agoricNames, but since the design details, especially around maintenance, weren't entirely clear, we rolled it back before shipping Fast USDC

chore: do not write chain info to vstorage #11312

p.s. I see Upload CrossChain Registry Contract V1 Sep 14 2023... looks like there's an owner who has edit rights:

    // Only contract governor can call denom alias CRUD operations
    let is_owner = is_owner(deps.as_ref(), &sender);
    let is_global_admin = is_global_admin(deps.as_ref(), &sender);

    if !is_owner && !is_global_admin {
        return Err(ContractError::Unauthorized {});
    }

-- https://github.com/osmosis-labs/osmosis/blob/01357893442a4d52d20de1401988a865ace498d1/cosmwasm/contracts/crosschain-registry/src/execute.rs#L258C1-L264C6

anilhelvaci · 2025-05-19T15:45:09Z

34 commits... some include "WIP"... @anilhelvaci it looks like you plan to squash; that is: review by commit is not the way to go. Can you confirm?

Correct.

Interesting. How are changes to the registry governed?

Oh, the contract needs an owner address in its constructor.

chore(multichain): set pool config json via script feat(multichain): add osmosis pool to CI E2E Workflow fix(multichain): fix lint issues fix(multichain): fix multichain workflow Invalid input chore(multichain): change osmosis_pool input type on CI workflow fix(multichain): fix Unable to use a TTY issue fix(multichain): fix timing issue when creating pool fix(multichain): remove kubectl tty flag refactor(multichain): reduce duplicated code squash multiple commits

Refs: #8863 fix: lint fixes

Refs: #8863

…ables Additional optimizations were included to handle json objects A new shell script replaced the use of a makefile to set the osmosis XCS state Rebase conflicts handled

Conflicts handled

ref: #8863

Refs: #8863

dckc

misc notes so far...

dckc · 2025-05-19T15:48:11Z

.github/workflows/multichain-e2e.yml

+      github.event_name == 'workflow_call' ||
+      github.event_name == 'pull_request' ||
+      (github.event_name == 'workflow_dispatch' && inputs.test_type == 'swap-on-osmosis')
+    uses: ./.github/workflows/multichain-e2e-template.yml


so this starts another set of nodes? How long does that take? It appears that landing this would slow down ci for all PRs going forward. Can you estimate how much?

dckc · 2025-05-21T15:25:49Z

multichain-testing/Makefile

+###############################################################################
+###                          Osmosis Setup                                   ###
+###############################################################################


I'm not a fan of ascii art comments, but I suppose that predates this PR.

Suggested change

###############################################################################

### Osmosis Setup ###

###############################################################################

###############################################################################

### Osmosis Setup ###

###############################################################################

dckc · 2025-05-21T15:30:58Z

multichain-testing/scripts/create-osmosis-pool.sh

+  balance_1=${balance_1:-0}
+  balance_2=${balance_2:-0}
+
+  if [[ "$balance_1" -ge "$TOKEN_IN_AMOUNT" && "$balance_2" -ge "$TOKEN_OUT_AMOUNT" ]]; then


The tedious best practice of quoting shell variable reference is followed here but not for $OSMOSIS_ADDRESS above. Run shellcheck, please.

What's the cost of adding shell check linting in ci here? I think we have it in some packages / repositories.

I sure don't like adding hundreds of lines of shell code. I'd rather see .js, perhaps using execa.

dckc · 2025-05-21T15:37:51Z

multichain-testing/scripts/download-wasm-artifacts.sh

@@ -0,0 +1,26 @@
+#!/bin/bash
+
+# Usage: ./download_specific_files.sh <owner> <repo> <branch> <folder> <destination> <file1> <file2> ...


For all shell scripts, add set -euo pipefail or at least set -e.

cf. a3p doctor
https://github.com/Agoric/agoric-3-proposals/blob/main/packages/synthetic-chain/src/cli/doctor.ts#L15C77-L15C94

dckc · 2025-05-21T15:41:32Z

multichain-testing/README.md

@@ -136,6 +136,13 @@ make provision-smart-wallet ADDR=$ADDR
 kubectl exec -i noblelocal-genesis-0 -c validator -- nobled query interchain-accounts host params | jq
 ```

+## Optional Features


Don't make folks read the details to know whether the section is relevant:

Suggested change

## Optional Features

## Osmosis Pool (optional)

dckc · 2025-05-21T15:59:38Z

multichain-testing/test/xcs-swap-anything/Makefile

+TX_FLAGS=--from $(OSMOSIS_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1uosmo --gas-adjustment 1.3 --yes --chain-id osmosislocal
+AGD_TX_FLAGS=--from $(AGORIC_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1ubld --gas-adjustment 1.3 --yes --chain-id agoriclocal


suggest: make it clearer which parts are the same by moving them to the beginning and making the variable names the same length:

Suggested change

TX_FLAGS=--from $(OSMOSIS_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1uosmo --gas-adjustment 1.3 --yes --chain-id osmosislocal

AGD_TX_FLAGS=--from $(AGORIC_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1ubld --gas-adjustment 1.3 --yes --chain-id agoriclocal

OSM_TX_FLAGS=--keyring-backend=test --gas=auto --gas-adjustment 1.3 --yes --from $(OSMOSIS_GENESIS_ADDRESS) --gas-prices 0.1uosmo --chain-id osmosislocal

AGD_TX_FLAGS=--keyring-backend=test --gas=auto --gas-adjustment 1.3 --yes --from $(AGORIC_GENESIS_ADDRESS) --gas-prices 0.1ubld --chain-id agoriclocal

If this were in .js, it would be straightforward to reify the authority wielded by these accounts as js objects. I haven't landed it yet, but some work in progress...

agoric-sdk/multichain-testing/tools/agd-lib.js

Line 179 in 9420b74

export const makeAgdAcct = (

Refs: #8863

dckc

PLEASE EXCUSE THE NOISE

I'm trying out some different tools; I don't really know how they work. But I don't want to throw my work away, so I'm posting this review as is...

dckc · 2025-05-22T19:52:08Z

multichain-testing/config.xcs-swap-anything.yaml

@@ -0,0 +1,125 @@
+name: agoric-multichain-testing


Something sped up multichain-testing a bunch recently... I wonder whether this config consistent with it... ah; yes, it's completely orthogonal:

feat: add and use chain.scripts.chainRpcReady hyperweb-io/starship#688

dckc · 2025-05-22T19:59:06Z

multichain-testing/scripts/create-osmosis-pool.sh

+AGORIC_WALLET="test1"
+AGORIC_ADDRESS=$(agoric-exec keys show ${AGORIC_WALLET} -a)
+OSMOSIS_WALLET="test1"


suggest: only use SHOUTING for variables exported to other processes via env; i.e. to say "hey! this variable is dynamically scoped! take extra care!". For ordinary variables that stay within the script, no need for shouting.

dckc · 2025-05-22T21:33:17Z

multichain-testing/test/xcs-swap-anything/README.md

+make create-osmosis-pool
+
+## Configure the osmosis registries
+cd cd agoric-sdk/multichain-testing/test/xcs-swap-anything


Suggested change

cd cd agoric-sdk/multichain-testing/test/xcs-swap-anything

cd agoric-sdk/multichain-testing/test/xcs-swap-anything

dckc · 2025-05-22T21:41:44Z

multichain-testing/test/xcs-swap-anything/swap-anything.test.ts

+    useChain,
+  } = t.context;
+
+  const { client, address } = await createFundedWalletAndClient(


This seems to use ambient access to a source of randomness. I suggest injecting it explicitly.

Distressing proliferation of ambient authority #2160

client seems to be based on ambient access to the network too: createFundedWalletAndClient uses SigningStargateClient.connectWithSigner

dckc · 2025-05-22T21:52:29Z

multichain-testing/test/xcs-swap-anything/swap-anything.test.ts

+    t.log(`Provisioned Agoric smart wallet for ${agoricAddr}`);
+
+    const apiUrl = await useChain('agoric').getRestEndpoint();
+    const queryClient = makeQueryClient(apiUrl);


makeQueryClient uses ambient access to fetch.

dckc · 2025-05-22T22:02:36Z

multichain-testing/test/xcs-swap-anything/swap-anything.test.ts

+      'ubld',
+    );
+
+    const { swapAddress } = await getXcsContractsAddress();


a function of 0 arguments can only return a constant, unless it uses ambient authority.

This one is using $.

dckc · 2025-05-22T22:02:57Z

multichain-testing/test/xcs-swap-anything/swap-anything.test.ts

+    const { swapAddress } = await getXcsContractsAddress();
+
+    // Send swap offer
+    const makeAccountOfferId = `swap-ubld-uosmo-${Date.now()}`;


ambient access to the clock

dckc · 2025-05-24T02:46:25Z

multichain-testing/test/xcs-swap-anything/helpers.ts

PLEASE EXCUSE THE NOISE pt 2

anilhelvaci force-pushed the swap-it-async-flow branch from c006458 to de48a00 Compare April 21, 2025 06:27

dckc mentioned this pull request Apr 25, 2025

docs: orchestration contract package skeleton #11323

Draft

anilhelvaci force-pushed the swap-it-async-flow branch from ea69e65 to 9da7798 Compare April 28, 2025 11:52

Jorge-Lopes force-pushed the swap-it-async-flow branch from 60b3e9e to fb8e9ae Compare May 6, 2025 14:22

anilhelvaci added the force:integration Force integration tests to run on PR label May 9, 2025

anilhelvaci marked this pull request as ready for review May 9, 2025 12:49

anilhelvaci requested a review from a team as a code owner May 9, 2025 12:49

anilhelvaci requested a review from AgoricTriage May 9, 2025 12:49

anilhelvaci assigned LuqiPan May 9, 2025

anilhelvaci mentioned this pull request May 14, 2025

chore: implement parseInboundTransfer #11375

Merged

anilhelvaci force-pushed the swap-it-async-flow branch from 86ba674 to 7b5170e Compare May 14, 2025 11:56

dckc self-requested a review May 14, 2025 17:23

Jorge-Lopes and others added 13 commits May 20, 2025 17:06

fix(multichain): change agoric-osmosis channel id

18df14e

chore(multichain): create pool by default and update names for clarity

8790505

fix(multichain): handle inconsistent channel-id

83172f9

chore(multichain): remove temporary pool config file

6c093a0

chore(multichain): do not create osmosis pool by default

9c209ff

fix(multichain): derive ibc-denom

0393e5b

chore: store and instantiate Osmosis XCS contracts

04ddc4a

Refs: #8863 fix: lint fixes

feat(multichain): store osmosis pool on swaprouter

47d6561

chore(orchestration): create scaffold for swap anything asyncFlow

3212873

chore: WIP swap anything

7b9fc96

Refs: #8863

chore: unit test for happy path

ad804ba

Refs: #8863

chore: init multichain testing work

38e87ed

Refs: #8863

anilhelvaci and others added 19 commits May 20, 2025 17:09

chore: WIP basic use case multichain testing

915fd4d

Refs: #8863

chore: improve assertions, add README

2ababc8

Refs: #8863

chore: integrate address hooks, unit tests partially complete

12eacb8

Refs: #8863

chore: add multichain tests for swapAnything with address hooks

8370136

Refs: #8863

test: Automate XCS setup via ava test

f66cf1a

chore: improve error handling

003f01c

fix: style fixes from prettier

26a0be1

Refs: #8863

chore: WIP ci setup

8604dbf

Refs: #8863

chore: remove address hooks related code

af18e2b

Refs: #8863

chore: address account sequence mismatch

9f1551d

chore: refactor osmosis swap test and script to remove hardcoded vari…

c4875a7

…ables Additional optimizations were included to handle json objects A new shell script replaced the use of a makefile to set the osmosis XCS state Rebase conflicts handled

test: verify osmosis xcs state

fa9b1eb

Conflicts handled

test: swap BLD for OSMO with receiver on CosmosHub

aa7e051

ref: #8863

chore: add helper functions to dedicated file

ab19545

test: swap OSMO for BLD with receiver on Agoric

c01fdd5

chore: implement failing IBC msg test case

3a6e4f2

Refs: #8863

fix: format fixes

e370ba3

chore: invitation name and tests updated, tests use dynamic xcs address

fc04b90

Refs: #8863

fix: exclude swap-anything from the main ava config

1e3d148

Refs: #8863

Jorge-Lopes force-pushed the swap-it-async-flow branch from 18f06c0 to 1e3d148 Compare May 20, 2025 16:09

dckc reviewed May 21, 2025

View reviewed changes

chore: bring back address hooks

e962549

Refs: #8863

anilhelvaci mentioned this pull request May 23, 2025

chore: handle pfm of transfer with memo #11410

Open

dckc reviewed May 24, 2025

View reviewed changes

multichain-testing/test/xcs-swap-anything/helpers.ts

Copy link

Member

dckc May 24, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

PLEASE EXCUSE THE NOISE pt 2

	const buildXCSMemo = swapInfo => {
	const memo = {
	wasm: {
	contract: swapInfo.destAddr,
	msg: {
	osmosis_swap: {
	output_denom: swapInfo.outDenom,
	slippage: {
	twap: {
	window_seconds: swapInfo.slippage.windowSeconds,
	slippage_percentage: swapInfo.slippage.slippagePercentage,
	},
	},
	receiver: swapInfo.receiverAddr,
	on_failed_delivery: swapInfo.onFailedDelivery,
	nextMemo: swapInfo.nextMemo,
	},
	},
	},
	};

		@@ -0,0 +1,26 @@
		#!/bin/bash

		# Usage: ./download_specific_files.sh <owner> <repo> <branch> <folder> <destination> <file1> <file2> ...

		TX_FLAGS=--from $(OSMOSIS_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1uosmo --gas-adjustment 1.3 --yes --chain-id osmosislocal
		AGD_TX_FLAGS=--from $(AGORIC_GENESIS_ADDRESS) --keyring-backend=test --gas=auto --gas-prices 0.1ubld --gas-adjustment 1.3 --yes --chain-id agoriclocal

	cd cd agoric-sdk/multichain-testing/test/xcs-swap-anything
	cd agoric-sdk/multichain-testing/test/xcs-swap-anything

chore(orchestration): create scaffold for swap anything asyncFlow #11304

Are you sure you want to change the base?

chore(orchestration): create scaffold for swap anything asyncFlow #11304

Uh oh!

Conversation

anilhelvaci commented Apr 21, 2025 • edited by Jorge-Lopes Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Security Considerations

Documentation Considerations

Testing Considerations

Upgrade Considerations

Uh oh!

cloudflare-workers-and-pages bot commented Apr 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Deploying agoric-sdk with Cloudflare Pages

Uh oh!

dckc commented May 19, 2025

Uh oh!

dckc commented May 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

anilhelvaci commented May 19, 2025

Uh oh!

dckc left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dckc left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

anilhelvaci commented Apr 21, 2025 •

edited by Jorge-Lopes

Loading

cloudflare-workers-and-pages bot commented Apr 21, 2025 •

edited

Loading

dckc commented May 19, 2025 •

edited

Loading