counterparty-guard
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
scripts/fetch_counterparty.pyto collect data. This script uses standard libraries and performs targeted requests to official government endpoints without executing arbitrary code. - [DATA_EXFILTRATION]: The skill performs outbound network requests to
egrul.nalog.ru,pb.nalog.ru, andbo.nalog.gov.ruto retrieve counterparty information. This behavior is documented and represents the primary functionality of the skill using official government services. - [PROMPT_INJECTION]: The skill processes untrusted data from external websites (aggregators and government portals), which is an indirect prompt injection surface. However, the skill provides specific 'Safety-floor' instructions to the agent to treat all scraped content as data rather than commands, mitigating the risk.
- [EXTERNAL_DOWNLOADS]: The skill identifies and references several well-known business data aggregators (e.g., checko.ru, saby.ru) for browser-based data collection. These references are used for information gathering and do not involve downloading or executing untrusted executable content.
Audit Metadata