jobs-evaluate

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). These are raw GitHub links to a PowerShell and shell installer script that the skill instructs you to fetch and execute directly (curl|bash, irm|iex); running remote .sh/.ps1 from an unknown or unverified repo is high-risk and commonly used to distribute malware unless the GitHub org/repo is trusted and you audit the scripts first.

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

• Third-party content exposure detected (high risk: 0.75). The skill fetches a cached “offer” (from a prior /jobs-search), and that offer text is outsider-authored job listing content that the LLM will read at runtime to evaluate and generate the rationale/dimensions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill can install a required runtime dependency by fetching and executing remote install scripts (curl -fsSL https://raw.githubusercontent.com/solid-company/solid-jobs-skills/main/scripts/install-sjctl.sh | bash and irm https://raw.githubusercontent.com/solid-company/solid-jobs-skills/main/scripts/install-sjctl.ps1 | iex), which downloads and executes remote code if sjctl is not present.