Subagents Explore Only
Turns every subagent into a read-only explorer. A SubagentStart hook injects an instruction into each subagent's context (built-in, custom and plugin agents alike) before its first prompt: investigate with read-only tools, never edit, create or delete files, never run state-changing commands, and report findings back so the main agent makes every change. The main conversation is not affected. This is an instruction, not enforcement: install the security/subagent-read-only-guard hook alongside it to block writes from subagents.
npx claude-code-templates@latest --setting hooks/subagents-explore-only Scanned by NVIDIA SkillSpector
Automated static analysis (no LLM) · v2.12.0 · 2026-10-17
SkillSpector looks for patterns such as shell commands, credential access, network calls or unpinned dependencies. An observation is something worth reading in the files before you install, not a confirmed vulnerability. Components that run code locally, like hooks, MCP servers and security tooling, naturally show more of them.
No observations.
Learn more about NVIDIA SkillSpector.