linkedin-content

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The content includes explicit instructions and admissions enabling privacy-invasive data collection and evasion (scraping that "bypass[es] anti-bot measures", de-anonymization of visitors) plus deliberate deliverability-evasion tactics (rotating secondary domains and many mailboxes), which together indicate intentional abusive behavior and high risk of misuse.

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

• Third-party content exposure detected (high risk: 0.75). ColdIQ MCP / API Marketplace are invoked at runtime to extract post engagement and signals (outsider-authored data from third-party marketplace/MCP), which the agent’s LLM then uses for drafting/routing.

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

• Potentially malicious external URL detected (high risk: 0.85). The skill explicitly instructs running remote code at runtime via "COLDIQ_API_KEY= npx -y @coldiq/mcp@latest", which fetches and executes an external package (and ties into ColdIQ endpoints such as https://api.coldiq.com) and therefore can directly introduce remote-controlled prompts or code execution.