backlink

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). This list includes a direct git clone of a personal/unknown GitHub repository (https://github.com/s87343472/backlink-pilot.git) that the skill instructs you to fetch and run, which is a security risk because executing code from an untrusted numeric/personal account may install malware or run arbitrary commands.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This code is an automated bulk backlink submission tool designed for mass/unattended directory postings with explicit evasion tactics (varying anchors, avoiding duplicate submissions, JS-eval form filling); it is an abuse-oriented spam/SEO manipulation tool though it contains no obvious credential-theft backdoor or remote-control implant.

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/batch-direct.mjs, the required runtime workflow fetches outsider-authored free text from each target site by opening site.url, taking a bb-browser snapshot, and then reading/parsing document text via bb-browser eval on document.body.innerText for confirmation.

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

• Potentially malicious external URL detected (high risk: 0.90). The Quick Start clones and then installs/runs code from a remote GitHub repo (git clone https://github.com/s87343472/backlink-pilot.git && npm install && node ...), which fetches and executes remote code at runtime.