grimoire-across

Warn

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructs the agent to run @grimoirelabs/cli using npx -y. This downloads and executes the package directly from the npm registry without version pinning or manual verification, creating a dependency on the integrity of the @grimoirelabs npm organization.- [EXTERNAL_DOWNLOADS]: The skill makes network requests to the Across Protocol public API at https://app.across.to/api/deposits/status to track the status of cross-chain transactions.- [COMMAND_EXECUTION]: The skill utilizes shell-based CLI commands to query blockchain bridge data, quotes, and routes.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Across Protocol API and CLI tool outputs into the agent context. While the current use case focuses on structured bridge data, this ingestion point represents a surface where a compromised external source could attempt to influence agent behavior. Boundary markers are currently absent in the provided instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Oct 16, 2026, 11:24 PM