Limit Order fixes

.codexignore

@@ -0,0 +1,4 @@
+.env
+.env.*
+**/*.env**/*.env.*
+agent/.env

.env.mainnet.example

@@ -1,9 +1,21 @@
 # Fork source
 MAINNET_RPC_URL=...
 
-# Deployer
+# Deployer (use DEPLOYER_PK directly or resolve it at runtime via agent/with-signer.mjs)
 DEPLOYER_PK=0x...
 
+# Optional signer resolution (used with agent/with-signer.mjs)
+# SIGNER_TYPE=env
+# PRIVATE_KEY=0x...
+# KEYSTORE_PATH=./keys/deployer.json
+# KEYSTORE_PASSWORD=...
+# KEYCHAIN_SERVICE=og-deployer
+# KEYCHAIN_ACCOUNT=deployer
+# VAULT_ADDR=https://vault.example.com
+# VAULT_TOKEN=...
+# VAULT_SECRET_PATH=secret/data/og-deployer
+# VAULT_SECRET_KEY=private_key
+
 # Optimistic Governor config
 OG_COLLATERAL=0xA0b86991c6218b36c1d19d4a2e9eb0ce3606eb48
 OG_BOND_AMOUNT=250000000

.gitignore

@@ -8,14 +8,17 @@ out/
 /broadcast/**/dry-run/
 /broadcast/*
 
-# Docs
-docs/
-
 # Dotenv file
 .env
 .env.mainnet
 .env.sepolia
+agent/.env
 
 # Anvil files
 anvil.log
-anvil.pid
+anvil.pid
+
+# Node modules
+node_modules
+frontend/node_modules
+agent/node_modules

AGENTS.md

@@ -1,10 +1,23 @@
 # Repository Guidelines
 
+## Documentation Hierarchy
+- Keep contribution guidance in both human docs and machine-readable instructions.
+- `AGENTS.md`: normative instructions for coding agents and automation.
+- `README.md`: architecture, intent, and operational context for humans.
+- `CONTRIBUTING.md`: shared contributor workflow and policy across the repo.
+- `skills/add-agent-commitment/SKILL.md`: reusable workflow for adding new agent/commitment combos.
+- Precedence when instructions conflict:
+1. Closest file to the changed code path wins.
+2. `AGENTS.md` is authoritative for agent behavior.
+3. Root-level files apply unless overridden by a closer area-level file.
+
 ## Project Structure & Module Organization
 - **`src/`**: Core Solidity contracts (e.g., `Counter.sol`). Keep new modules grouped by domain and include SPDX + pragma headers.
 - **`script/`**: Deployment and automation scripts (e.g., `Counter.s.sol`, `DeploySafeWithOptimisticGovernor.s.sol`). Favor reusable helpers and parameterize via environment variables.
 - **`test/`**: Forge tests using `forge-std`’s `Test` base. Mirror contract names (`<Name>.t.sol`) and co-locate fixtures with the subject under test.
 - **`lib/`**: External dependencies (currently `forge-std`) managed through Foundry.
+- **`agent/`**: Shared offchain runner, signer integrations, and reusable tooling.
+- **`agent-library/`**: Agent-specific implementations under `agent-library/agents/<name>/`.
 
 ## Build, Test, and Development Commands
 - `forge build`: Compile all contracts.
@@ -20,12 +33,22 @@
 - Keep files focused and small; prefer internal helpers over inline duplication.
 - Run `forge fmt` to enforce style; include SPDX identifiers and explicit visibility where practical.
 - Use descriptive variable names (avoid single letters outside of loop counters or hashes).
+- For Node.js code in `agent/` and `agent-library/`, keep modules small and isolate side effects at the edges.
+
+## Agent Locality Rule
+- New functionality for a specific agent must be implemented in that agent's own files under `agent-library/agents/<agent-name>/`.
+- Do not add agent-specific behavior to shared generalized files in `agent/src/lib/` or `agent/src/index.js`.
+- Shared generalized files should only change when:
+1. The change is required for multiple agents.
+2. The change fixes a bug in shared infrastructure.
+- If a pull request changes shared generalized agent files, include a brief cross-agent rationale in the PR description and link impacted agents.
 
 ## Testing Guidelines
 - Use `forge-std/Test` utilities for assertions, fuzzing (`testFuzz_*`), and logging.
 - Name tests with behavior-first patterns (`test_IncrementsCounter`, `testFuzz_SetNumberMaintainsState`).
 - Cover success, failure, and access-control paths; add revert expectation tests when changing critical flows.
 - When modifying gas-sensitive code, refresh `forge snapshot` and include notes in PRs.
+- For `agent-library` changes, run the relevant agent test/simulation scripts and note commands in the PR.
 
 ## Commit & Pull Request Guidelines
 - Write imperative, concise commit messages (e.g., "Add OG deployment script"); group related changes together.

CONTRIBUTING.md

@@ -0,0 +1,47 @@
+# Contributing
+
+This repository uses two kinds of contribution guidance:
+
+- `AGENTS.md` for machine-targeted, normative instructions.
+- `README.md` for human-oriented architecture and workflow context.
+
+## Precedence Rules
+
+When instructions conflict:
+
+1. The closest file to the code you are editing wins.
+2. `AGENTS.md` is authoritative for agent behavior.
+3. Root-level guidance applies when there is no closer override.
+
+## Directory Guidance
+
+- `src/`, `script/`, `test/`: Solidity contracts, scripts, and tests.
+- `agent/`: shared offchain runner and reusable agent infrastructure.
+- `agent-library/agents/<name>/`: agent-specific implementations.
+- `docs/`: operational and architecture docs.
+
+## Required Contributor Workflow
+
+1. Read relevant local docs before editing (`AGENTS.md`, `README.md`).
+2. Keep changes scoped to the correct area.
+Agent-specific behavior belongs in `agent-library/agents/<name>/`.
+Shared runner changes in `agent/` require cross-agent justification.
+3. Run the minimum required checks.
+Solidity changes: `forge fmt`, `forge test`.
+Agent changes: relevant module tests/simulations.
+4. In PRs, document:
+What changed and why.
+Tests run.
+Any config or environment variable impacts.
+
+## Agent Extension Policy
+
+New agent behavior must be added to that agent's own library files instead of shared generalized files.
+
+- Preferred location: `agent-library/agents/<agent-name>/agent.js` and related files in that directory.
+- Shared files such as `agent/src/lib/*` and `agent/src/index.js` should only change for multi-agent abstractions or shared bug fixes.
+- If shared files are changed, the PR must include:
+Why an agent-local implementation was insufficient.
+Which existing agents are impacted.
+
+See `docs/agent-extension-guidelines.md` for the decision framework and examples.

README.md

@@ -1,19 +1,23 @@
-## OG Deployer
+# Oya Commitments
 
-Command-line tooling for deploying onchain Commitments: a Safe with an Optimistic Governor module configured with natural language rules. Use it to spin up a new Safe, connect the Optimistic Governor module, and set the rules/bond/collateral in one script run.
+Oya Commitments are Safe-based commitments controlled by plain-language rules and enforced via an Optimistic Governor module. This repo contains the Solidity contracts, deployment scripts, an optional web UI, and an offchain agent scaffold.
 
-## What This Repo Does
+## Beta Disclaimer
 
-- Deploys a Safe and the Optimistic Governor module in one flow.
-- Encodes natural language rules into the module configuration.
-- Supports env-var overrides for Safe and OG parameters.
-- Uses Foundry for scripting, testing, and deployments.
+This is beta software provided “as is.” Use at your own risk. No guarantees of safety, correctness, or fitness for any purpose.
+
+## How It Works (At a Glance)
+
+- Write plain-language rules that define what the commitment may do.
+- Deploy a Safe wired to an Optimistic Governor module with those rules.
+- An agent or user proposes transactions via the module and posts a bond.
+- If no challenge occurs during the window, the proposal is executed by the Safe.
 
 ## Quick Start
 
 1. Install Foundry: https://book.getfoundry.sh/
-2. Set environment variables in your shell or `.env` file (load via `direnv` or `dotenvx` if desired).
-3. Run the deployment script.
+2. Set required environment variables (see `docs/deployment.md`).
+3. Run the deployment script:
 
 ```shell
 forge script script/DeploySafeWithOptimisticGovernor.s.sol:DeploySafeWithOptimisticGovernor \
@@ -22,165 +26,22 @@ forge script script/DeploySafeWithOptimisticGovernor.s.sol:DeploySafeWithOptimis
   --private-key <your_private_key>
 ```
 
-## Required Environment Variables
-
-- `DEPLOYER_PK`: Private key for the deployer.
-- `OG_COLLATERAL`: Address of the ERC20 collateral token.
-- `OG_BOND_AMOUNT`: Bond amount for challenges.
-- `OG_RULES`: Natural language rules for the commitment.
-
-## Optional Overrides
-
-- `SAFE_SALT_NONCE`, `SAFE_THRESHOLD`, `SAFE_OWNERS`
-- `OG_SALT_NONCE`, `OG_CHALLENGE_PERIOD`, `OG_RULES_URI`
-- `OG_MASTER_COPY`, `SAFE_SINGLETON`, `SAFE_FALLBACK_HANDLER`
-- `MODULE_PROXY_FACTORY`
+## Documentation
 
-## Example `.env`
+- Contribution workflow and policy: `CONTRIBUTING.md`
+- Skill for new agent/commitment combos: `skills/add-agent-commitment/SKILL.md`
+- Deployment and configuration: `docs/deployment.md`
+- Signer options and `with-signer` helper: `docs/signers.md`
+- Offchain agent usage: `docs/agent.md`
+- Agent extension decision rules: `docs/agent-extension-guidelines.md`
+- Web frontend: `docs/frontend.md`
+- Testing and common commands: `docs/testing.md`
 
-```ini
-# Required
-DEPLOYER_PK=0xabc123...
-OG_COLLATERAL=0x1111111111111111111111111111111111111111
-OG_BOND_AMOUNT=250000000
-OG_RULES="Any assets deposited in this Commitment may be transferred back to the depositor before January 15th, 2026 (12:00AM PST). After the deadline, assets may only be transferred to jdshutt.eth. If a third party is initiating the transfer after the deadline, they may take a 10% cut of the assets being transferred as a fee."
-
-# Safe overrides
-SAFE_OWNERS=0x2222222222222222222222222222222222222222,0x3333333333333333333333333333333333333333
-SAFE_THRESHOLD=2
-SAFE_SALT_NONCE=12345
-
-# Optimistic Governor overrides
-OG_CHALLENGE_PERIOD=604800
-OG_RULES_URI=ipfs://bafy...
-OG_SALT_NONCE=67890
-
-# Optional factory / master copy overrides
-MODULE_PROXY_FACTORY=0x4444444444444444444444444444444444444444
-OG_MASTER_COPY=0x5555555555555555555555555555555555555555
-SAFE_SINGLETON=0x6666666666666666666666666666666666666666
-SAFE_FALLBACK_HANDLER=0x7777777777777777777777777777777777777777
-```
+## Repo Layout
 
-## Common Commands
-
-```shell
-forge build
-forge test
-forge fmt
-```
-
-## Local Testing
-
-Dry-run (no broadcast):
-
-```shell
-anvil
-forge script script/DeploySafeWithOptimisticGovernor.s.sol:DeploySafeWithOptimisticGovernor \
-  --rpc-url http://127.0.0.1:8545 \
-  --private-key <your_private_key>
-```
-
-Broadcast on Anvil:
-
-```shell
-anvil
-forge script script/DeploySafeWithOptimisticGovernor.s.sol:DeploySafeWithOptimisticGovernor \
-  --rpc-url http://127.0.0.1:8545 \
-  --broadcast \
-  --private-key <your_private_key>
-```
-
-## Propose & Execute Transfers
-
-Propose a transfer (posts the UMA bond via the Optimistic Governor):
-
-```shell
-export PROPOSER_PK=<private_key>
-export OG_MODULE=<optimistic_governor_module>
-export TRANSFER_ASSET=<erc20_token_address>
-export TRANSFER_AMOUNT=<amount_in_token_units>
-export TRANSFER_DESTINATION=<recipient_address>
-
-forge script script/ProposeCommitmentTransfer.s.sol:ProposeCommitmentTransfer \
-  --rpc-url <your_rpc_url> \
-  --broadcast \
-  --private-key $PROPOSER_PK
-```
-
-Execute a proposal after it passes:
-
-```shell
-export EXECUTOR_PK=<private_key>
-export OG_MODULE=<optimistic_governor_module>
-export PROPOSAL_HASH=<proposal_hash>
-export TRANSFER_ASSET=<erc20_token_address>
-export TRANSFER_AMOUNT=<amount_in_token_units>
-export TRANSFER_DESTINATION=<recipient_address>
-
-forge script script/ExecuteCommitmentTransfer.s.sol:ExecuteCommitmentTransfer \
-  --rpc-url <your_rpc_url> \
-  --broadcast \
-  --private-key $EXECUTOR_PK
-```
-
-Optional overrides:
-
-- `TRANSFER_OPERATION` (default `0` for `CALL`)
-- `TRANSFER_VALUE` (default `0`)
-
-### Anvil Test Key + USDC Funding (Fork)
-
-Start Anvil with the default test mnemonic and grab one of the printed private keys:
-
-```shell
-anvil --fork-url $MAINNET_RPC_URL --mnemonic "test test test test test test test test test test test junk"
-```
-
-Fund the test account with USDC by impersonating a whale on the fork:
-
-```shell
-cast rpc anvil_impersonateAccount <whale_address>
-cast rpc anvil_setBalance <whale_address> 0x3635C9ADC5DEA00000
-cast send <usdc_contract> "transfer(address,uint256)" <your_account> <amount> --from <whale_address>
-cast rpc anvil_stopImpersonatingAccount <whale_address>
-```
-
-## Network Env Files
-
-You can keep per-network env files and load them with a tool like `dotenvx` or `direnv`.
-
-Mainnet fork example (`.env.mainnet`):
-
-```ini
-MAINNET_RPC_URL=...
-DEPLOYER_PK=0x...
-OG_COLLATERAL=0xA0b86991c6218b36c1d19d4a2e9eb0ce3606eb48
-OG_BOND_AMOUNT=250000000
-OG_RULES="Any assets deposited in this Commitment may be transferred back to the depositor before January 15th, 2026 (12:00AM PST). After the deadline, assets may only be transferred to jdshutt.eth. If a third party is initiating the transfer after the deadline, they may take a 10% cut of the assets being transferred as a fee."
-OG_IDENTIFIER_STR=ASSERT_TRUTH2
-```
-
-A ready-to-edit template is available at `.env.mainnet.example`.
-
-Sepolia example (`.env.sepolia`):
-
-```ini
-SEPOLIA_RPC_URL=...
-DEPLOYER_PK=0x...
-SAFE_SINGLETON=0x...
-SAFE_PROXY_FACTORY=0x...
-SAFE_FALLBACK_HANDLER=0x...
-OG_MASTER_COPY=0x...
-OG_COLLATERAL=0x...
-OG_BOND_AMOUNT=...
-OG_RULES="Any assets deposited in this Commitment may be transferred back to the depositor before January 15th, 2026 (12:00AM PST). After the deadline, assets may only be transferred to jdshutt.eth. If a third party is initiating the transfer after the deadline, they may take a 10% cut of the assets being transferred as a fee."
-```
-
-Load the file before running the script:
-
-```shell
-dotenvx run -f .env.mainnet -- forge script script/DeploySafeWithOptimisticGovernor.s.sol:DeploySafeWithOptimisticGovernor \
-  --rpc-url $MAINNET_RPC_URL \
-  --private-key $DEPLOYER_PK
-```
+- `src/` Solidity contracts
+- `script/` Foundry deployment and ops scripts
+- `test/` Foundry tests
+- `agent/` Offchain agent scaffold
+- `frontend/` Web UI
+- `lib/` External dependencies (Foundry)

agent-library/AGENTS.md

@@ -0,0 +1,24 @@
+# Agent Library Guidelines
+
+## Scope
+
+This file applies to `agent-library/` and `agent-library/agents/*`.
+
+## Purpose
+
+This directory is the home for agent-specific behavior and commitment-specific decision logic.
+
+## Rules
+
+- Each agent lives under `agent-library/agents/<agent-name>/`.
+- Keep commitment logic, prompt strategy, and behavior specialization in that agent's local files.
+- Prefer adding new modules over branching shared runner code for one-off behavior.
+
+## Locality Rule
+
+When creating a new agent, place functionality in that agent's own files (`agent.js` and adjacent module files). Do not implement single-agent behavior in shared generalized files under `agent/src/lib/` or `agent/src/index.js`.
+
+## Validation
+
+- Run module-specific tests/simulations for the changed agent.
+- Document test commands in the PR description.