ai-repo-setup

Pass

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses standard system commands like ln -s to maintain compatibility across agent tools and recommends chmod +x for configuring local hook scripts, both of which are common for repository initialization.
  • [EXTERNAL_DOWNLOADS]: The skill instructs agents to install well-known and reputable developer packages such as Lefthook, Husky, Biome, and Oxlint from standard registries.
  • [PROMPT_INJECTION]: The skill provides a framework for analyzing repository structure and requirements, which creates a surface for indirect prompt injection from untrusted code. Ingestion points include project metadata and functional requirements; boundary markers are defined using status-based markdown headers; capabilities include file-system modifications and local command execution; no explicit sanitization is described for the extracted content.
  • [DYNAMIC_EXECUTION]: The skill generates templates for local bash scripts and JavaScript plugins to enforce repository rules deterministically, enhancing the reliability of agent tool usage without executing remote unverified code.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 16, 2026, 01:33 PM