caddy
Fail
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: HIGHPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill documentation suggests running commands with elevated privileges ("sudo caddy start") to resolve permission issues when binding to restricted ports like port 80.
- [COMMAND_EXECUTION]: The skill instructs the agent to use user-provided names exactly as specified in shell commands, creating a command injection vulnerability. If a user provides a name containing shell metacharacters, it will be interpolated directly into curl commands. Evidence: curl -sf -X DELETE "http://localhost:2019/id/proxy_NAME".
- [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerable surface for processing untrusted data. 1. Ingestion points: User-provided proxy names. 2. Boundary markers: Absent. 3. Capability inventory: Shell-based curl and lsof commands. 4. Sanitization: Absent for user-provided names, although present for auto-generated names.
Recommendations
- AI detected serious security threats
Audit Metadata