amazon-top-reviews

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.90). The skill downloads and immediately executes a remote shell script from a GitHub raw URL (raw.githubusercontent.com) pointing to an external repository (FinpeakInc/frevana-cli-releases), which is effectively running untrusted remote code and is therefore high risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This code intentionally downloads and executes an unverified remote setup script before every run, creating a high-risk supply-chain / remote code execution vector that could be used to install backdoors or exfiltrate data (the tool also accesses the local Frevana daemon / Chrome session which increases the potential impact).

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

• Third-party content exposure detected (high risk: 0.85). Yes—scripts/get_amazon_top_reviews.sh passes the user-supplied Amazon product URL to the local frevana_ask tool, which (via the Chrome extension/daemon) fetches outsider-authored Amazon review text and returns it as readable output ($RESULT_FILE), which is then read and included in the agent’s context via cat "$RESULT_FILE".

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

• Potentially malicious external URL detected (high risk: 1.00). The skill's setup wrapper (scripts/setup.sh) downloads and execs a remote shell script at runtime from https://raw.githubusercontent.com/FinpeakInc/frevana-cli-releases/refs/heads/main/skills/frevana/scripts/setup.sh, which executes remote code and is required before tool use.