yield-opportunities
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's research scripts ingest user-supplied parameters (token names, blockchain identifiers, and export file paths) which are used directly in shell commands and file system operations.
- Ingestion points: Command-line arguments in
scripts/compare-protocols.shandscripts/fetch-defillama-yields.py(e.g.,--token,--chain,--export). - Boundary markers: None identified; parameters are interpolated directly into strings.
- Capability inventory: The scripts perform network operations (
curl,requests.get) and file-write operations (echo > "$EXPORT_FILE"). - Sanitization: Very basic; the shell script converts token names to uppercase but does not validate file paths or prevent directory traversal/overwriting of sensitive files.
- [EXTERNAL_DOWNLOADS]: The skill interacts with external APIs to fetch financial data.
- Fetches yield pool data from DefiLlama's official API (
yields.llama.fi), a well-known service in the DeFi ecosystem. - Documentation references various third-party DeFi protocol dashboards (Aave, Beefy, Yearn) and vendor-owned repositories (
github.com/kukapay/defi-yields-mcp). - [COMMAND_EXECUTION]: The skill includes a shell script (
scripts/compare-protocols.sh) designed to be executed by the agent. - The script executes
curlto interact with remote APIs andjqfor data processing. - The output formatting uses
printfandecho -eon data retrieved from external sources.
Audit Metadata