swap-execute
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download the Foundry installation script from
foundry.paradigm.xyz, which is the official distribution point for the well-known Ethereum development toolkit. - [REMOTE_CODE_EXECUTION]: The documentation includes a setup step that pipes the downloaded Foundry script directly to the shell (
curl ... | bash). The skill explicitly notes this as a standard practice that carries supply chain risks. - [COMMAND_EXECUTION]: The skill's primary function involves generating and executing shell commands using Foundry's
casttool to simulate (cast call) and broadcast (cast send) blockchain transactions. - [INDIRECT_PROMPT_INJECTION]: The skill processes structured JSON data from a preceding
swap-buildstep to populate transaction fields. - Ingestion points: Reads transaction data (calldata, addresses, values) from the
swap-buildoutput inSKILL.md. - Boundary markers: Includes a mandatory 'Step 3: Confirm Execution' where the agent must present transaction details and wait for an explicit "yes" from the user before proceeding.
- Capability inventory: Execution of shell commands via
castand potential execution of JavaScript snippets viaethers.jsfallback. - Sanitization: Instructs the agent to validate input fields, simulate the transaction to check for reverts, and verify the router address against a known hardcoded value.
Audit Metadata