linkedin-content

Pass

Audited by Gen Agent Trust Hub on Aug 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references external resources for installation and configuration, specifically targeting github.com/inference-sh and github.com/belt-sh repositories.
  • [COMMAND_EXECUTION]: Utilizes the belt CLI tool to perform network-based operations such as content research and cross-platform posting to Twitter.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface where external or user-provided content is interpolated into shell commands.
  • Ingestion points: User-provided inputs for LinkedIn post generation and research queries within SKILL.md.
  • Boundary markers: Lacks explicit delimiters or specific instructions to sanitize or ignore embedded control characters in input text.
  • Capability inventory: The skill has access to the belt CLI tool, which is capable of network communication and interacting with third-party APIs.
  • Sanitization: The provided command examples use simple JSON quoting which may be bypassed by crafted user inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 8, 2026, 08:30 PM
Security Audit — agent-trust-hub — linkedin-content