podkop-openwrt-guide

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

• Third-party content exposure detected (high risk: 0.90). The skill's workflow and references (references/operations.md, references/repo-map.md, and SKILL.md) explicitly show runtime fetching of untrusted external lists and files (e.g., install.sh pulling raw.githubusercontent.com, remote_domain_lists/remote_subnet_lists and COMMUNITY_SERVICES URLs in podkop/files/usr/lib/constants.sh) which are parsed and used to generate routing rules/outbounds via podkop list_update and rulesets.sh, meaning third‑party content directly influences tool actions and decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs fetching and executing a remote install script at runtime via curl -fsSL "https://raw.githubusercontent.com/itdoginfo/podkop/${PODKOP_REF}/install.sh" and then running sh /tmp/podkop-install.sh, which downloads and executes remote code.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). This skill explicitly instructs the agent to install/update/migrate and modify system-level configuration and services on an OpenWrt device (e.g., /etc/config/podkop, dnsmasq, nft/tproxy, sing-box and service lifecycle), which entails privileged changes to the machine state.