agentwallet
Pass
Audited by Gen Agent Trust Hub on Mar 21, 2026
Risk Level: SAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill manages an API token and wallet configuration stored in a local file at ~/.agentwallet/config.json. It includes explicit instructions for the agent to set restrictive file permissions (chmod 600) and to prevent the token from being leaked in logs or conversation text.
- [EXTERNAL_DOWNLOADS]: The skill communicates with the frames.ag domain via curl to perform wallet actions, check balances, and retrieve statistics.
- [PROMPT_INJECTION]: The skill processes data from external APIs, creating a surface for indirect prompt injection.
- Ingestion points: API responses from frames.ag (e.g., stats, activity, pulse).
- Boundary markers: The skill instructions contain a specific warning to treat API response bodies as untrusted data.
- Capability inventory: The skill has the capability to perform financial transfers and contract calls.
- Sanitization: Relies on explicit instructions for the agent to ignore any commands embedded in processed API data.
Audit Metadata