hardening
Warn
Audited by Snyk on Nov 22, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (low risk: 0.10). The required Review workflow runs read-only commands on the user's own hosts, and the auth-log checks (
last,lastlog,journalctl -u sshd) can surface attacker-chosen strings such as SSH usernames from unauthenticated connection attempts, but the workflow neither monitors an outsider-writable queue nor selects outsider content.
MEDIUM W013: Attempt to modify system services in skill instructions.
- Attempt to modify system services in skill instructions detected (low risk: 0.10). Gate 1 is met because proposal.md gives concrete state-changing commands (
setenforce 0,systemctl reload sshd,firewall-cmd --timeout,rndc reconfig,postfix reload), but Gate 2 fails because no explicit unrecoverable loss of operational state is tied to an available operation: the irreversible vault lock prevents deletion rather than losing state, and the key-deletion reference is not stated to destroy retained data.
Issues (2)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
W013
MEDIUMAttempt to modify system services in skill instructions.
Audit Metadata