The Agent Skills Directory

Unverifiable Dependencies & Remote Code Execution (CRITICAL): The skill promotes the command curl -fsSL https://cli.inference.sh | sh for installation. This method downloads and executes code from a remote server without verification or integrity checks, which can lead to complete system compromise. This finding is confirmed by the automated security scan.\n- External Downloads (HIGH): The skill uses npx to install additional content from the inference-sh/skills repository. Neither the repository nor the organization are on the verified trusted list, posing a potential supply chain risk.\n- Indirect Prompt Injection (LOW): The skill retrieves external data using tavily/search-assistant and infsh/agent-browser (Ingestion points) and uses it to generate posts via x/post-create (Capability inventory). No boundary markers are used to isolate this untrusted data (Boundary markers), and there is no evidence of sanitization or validation of the fetched content (Sanitization).\n- Command Execution (MEDIUM): The skill relies on several bash commands via the infsh CLI to perform core functions, which increases the potential attack surface if parameters are not handled securely.

twitter-thread-creation