fluxa-agent-wallet
Pass
Audited by Gen Agent Trust Hub on Mar 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the
@fluxa-pay/fluxa-walletpackage and utilizes theawalpackage via npx, which are external dependencies from the npm registry. - [REMOTE_CODE_EXECUTION]: The use of
npx awal@2.0.3 x402to discover services involves downloading and executing code from a remote source at runtime. - [PROMPT_INJECTION]: The skill ingests external instructions from
https://clawpi-v2.vercel.app/api/skill.mdandmonetize.fluxapay.xyz. 1) Ingestion points: Markdown files and API responses from external domains. 2) Boundary markers: None. 3) Capability inventory: CLI tools for payments, payouts, and systemopencommand. 4) Sanitization: None. This surface allows for potential indirect prompt injection attacks. - [COMMAND_EXECUTION]: The skill executes various system commands through the
fluxa-walletCLI and accesses a sensitive local configuration file at~/.fluxa-ai-wallet-mcp/config.jsonfor Agent ID credential storage. It also uses the systemopencommand to initiate browser-based user authorizations.
Audit Metadata