customer-persona

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). They point to a custom CLI installer that asks users to download and execute a remote shell script and binaries hosted on dist.inference.sh with only an in-repo plaintext checksum for verification—piping a remote script to sh and trusting same-host checksums is a risky distribution pattern that could deliver malware if the domain, distribution server, or DNS are compromised.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md "Step 1: Research" explicitly instructs running commands like infsh app run tavily/search-assistant and infsh app run exa/search to query public web/search results, which the agent is expected to read and use to build personas—so untrusted third‑party content can materially influence subsequent decisions and actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The Quick Start explicitly instructs running "curl -fsSL https://cli.inference.sh | sh" (which downloads and executes a remote installer and retrieves binaries from dist.inference.sh) and the skill's examples rely on that CLI at runtime, so this URL is a required runtime dependency that executes remote code.