newsletter-curation

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). These are unknown third‑party domains serving an installable shell script (curl … | sh) from a nonstandard .sh host — executing remote scripts from unverified domains is a high‑risk vector for malware distribution.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md "Content Sourcing" section explicitly instructs running search apps (e.g., infsh app run tavily/search-assistant and exa/search) and even queries targeting public sites like "site:reddit.com" and social sources (Twitter, LinkedIn, HN), so the agent is expected to fetch and read untrusted, user-generated web content that can influence newsletter curation and subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's Quick Start instructs running curl -fsSL https://cli.inference.sh | sh which fetches and executes remote installer code at setup/runtime and the skill repeatedly relies on the resulting infsh CLI (infsh app run ...) as a required dependency, so https://cli.inference.sh is a high-risk runtime-executed URL.