publish-facebook-post

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). This is high risk: the skill downloads and immediately execs a raw script from a GitHub user/repo that isn’t an obviously trusted vendor, which is a direct remote-code-execution distribution vector.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This code repeatedly downloads and executes a remote setup script (from raw.githubusercontent.com) before every run, creating a high-risk supply-chain / remote-code-execution vector that could install backdoors or exfiltrate data; no explicit data-exfiltration code is present in the repo itself, but the remote-exec behavior enables it.

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

• Potentially malicious external URL detected (high risk: 1.00). The bundled scripts/setup.sh is invoked at runtime (scripts/publish_facebook_post.sh calls run_frevana_setup) and that wrapper downloads and execs remote code from https://raw.githubusercontent.com/FinpeakInc/frevana-cli-releases/refs/heads/main/skills/frevana/scripts/setup.sh, so remote content is fetched and executed as a required dependency.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly instructs the agent to download and execute a remote setup script (scripts/setup.sh) before every run and to run a local daemon, which can install services or change system configuration and therefore modifies the machine state and can compromise the host.