Environment variable access combined with network send.
- Code
- malicious.env_harvesting
- Location
- scripts/portfolio.js:152
Security audit
Security checks for vulnerabilities and agentic risk
The skill's functionality (an autonomous AI-only wallet) matches the code and runtime it installs, but several installation and persistence details (native binary/daemon, remote fetches, cron registration) raise meaningful risk and some internal inconsistencies that you should review before installing.
This skill is coherent with its stated purpose (an autonomous AI-only wallet) but carries elevated operational risk. Before installing: 1) Inspect scripts/provision.sh and any network download URLs to confirm exactly what is fetched and executed. 2) Verify the provenance of the native addon (core/aawp-core.node) and the on-chain hash check; do not run if you cannot validate the binary. 3) Understand where keys are stored (config/guardian.json, .agent-config/seed.enc) and consider using an isolated/sandboxed host or VM; do not install on a machine that already holds other sensitive keys or funds. 4) Require human approval for provisioning and wallet creation (SKILL.md suggests this) and avoid giving the skill broad autonomous permission until you audit the code and binary. 5) If you proceed, pin AAWP_WALLET immediately after creation, keep backups offline, and monitor the daemon's socket and cron registrations. If you want me to, I can extract and summarize the contents of scripts/provision.sh, core loader, or any other specific file for a deeper audit.
Detected: malicious.env_harvesting, suspicious.dangerous_exec, suspicious.potential_exfiltration