ssti-hunter

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill contains explicit, actionable exploitation instructions and payloads for achieving RCE, data exfiltration, reverse shells, sensitive-file/cloud-metadata access, and deploying web shells — clear malicious/abusive guidance.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s Phase 1–4 workflow ingests outsider-provided test context such as the target URL/brief and then drives probes by injecting attacker-controlled strings (e.g., report name / URL parameters / form fields) into the victim app’s template-sink to observe evaluated output.

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 contains explicit RCE payloads (command execution via popen/exec, file reads of /etc/passwd and /etc/shadow, writing a PHP webshell via writeFile, and reverse-shell/OOB commands) which instruct modifying or compromising a target system and could be used against the agent's host, so it poses a high risk of changing the machine state.