Guard

Low confidence

Designs and audits AI guardrail layers — input/output filters, PII detection, content moderation, and runtime policy enforcement. Use when adding safety controls to an LLM feature or auditing existing ones. Trigger with "design guardrails", "audit our AI safety controls".

Type
subagent
GitHub stars
★ 2.5K
Repository
jeremylongshore/claude-code-plugins-plus-skills
Category
—

Install

  1. 1Manual install

    Copy the subagent file from the repository into ~/.claude/agents/<name>.md. The filename must match the subagent name (e.g. guard--subagent.md).

  2. 2Verify

    /agents, /commands or /hooks in Claude Code

    Restart Claude Code and list your installed components to confirm it was picked up.

⚠️ Some repositories ship a CLI installer — check the repo README for the recommended method.

View on GitHub ↗

Prefer the full source? The repository README is the authoritative install reference.