hithink-finance-data
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to interact with the local environment by executing the
hithink-financecommand-line utility. It provides various subcommands for data management, includingdata init,sync,status, anddb query. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from external sources.
- Ingestion points: Data enters the agent's context through
db queryresults (local database content),data sync(remote data fetching), anddata init(importing local Parquet files). - Boundary markers: The skill explicitly instructs the agent that "SQL must be read-only" and provides usage constraints in the 'Boundaries' section of
SKILL.md. - Capability inventory: The skill can execute shell commands via the
hithink-financebinary, write data to the local file system using the--outputparameter, and perform network synchronization. - Sanitization: The skill relies on natural language instructions to enforce read-only behavior and does not demonstrate programmatic sanitization of the SQL queries or imported data.
Audit Metadata