building-skills

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The instructions guide the agent to use various CLI tools, including git for repository management and the amp utility for skill administration, such as amp skill add, amp clone, and amp skills import.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the acquisition of external content through Git repository cloning and skill importing from remote URLs. It also mentions using npx to fetch and run MCP servers from the npm registry.
  • [DYNAMIC_EXECUTION]: The skill describes how to configure MCP servers that run as standalone processes via shell commands (using command, args, and env fields), allowing for the runtime execution of tools defined within the skill frontmatter.
  • [INDIRECT_PROMPT_INJECTION]: This skill serves as a gateway for importing third-party skills, creating a surface for indirect prompt injection from processed data. It includes a mandatory review process for scripts and configurations, advises on using includeTools to filter capabilities to minimum requirements, and emphasizes auditing internet-sourced skills for unexpected network or credential access before they are loaded.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 03:21 AM