phpinfo-to-rce

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill is an explicit, intentional exploitation playbook that instructs discovery of exec-enabled PHP installs, locating upload vectors, creating/uploading a PHP webshell, and executing reverse shells — clearly malicious and designed for unauthorized RCE.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). SKILL.md’s required runtime workflow ingests outsider-authored free text only from user/attacker-controlled HTTP responses it fetches at execution time (e.g., curl ... https://$TARGET/info.php and related upload-vector endpoint pages), so the LLM is exposed to arbitrary external page/body text submitted via those target URLs.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). This skill explicitly instructs the agent to analyze phpinfo for exec availability, find an upload vector, upload a PHP webshell, execute system commands and even establish a reverse shell (including bypass techniques), which is a direct push to perform remote RCE and security-bypass actions.