hot-take-risk-check
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill contains instructions for the user to download and install a Model Context Protocol (MCP) server from the author's domain (
https://mcp.liftli.ai/mcp). This tool is used to extend the skill's capabilities with LinkedIn publishing features. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content (drafts) for analysis, creating a potential vector for indirect prompt injection where malicious text could attempt to manipulate the agent's output or tool usage.
- Ingestion points: User drafts are ingested and processed in the
Processsection of SKILL.md. - Boundary markers: Absent. The skill instructions do not implement delimiters or safety warnings to help the agent distinguish user content from its own instructions.
- Capability inventory: The skill is instructed to interact with the
strategist_briefingtool (SKILL.md) if available and refers to thelinkedin-publishskill for additional actions. - Sanitization: Absent. There are no instructions to sanitize or validate the user-supplied text before the agent evaluates it.
Audit Metadata