Agent Guard is a local-first guardrail for Claude Code, Codex, Git hooks, GitHub Actions, and direct shell use. It blocks common secret-exposure paths before a supported tool runs, redacts supported tool output, and scans changed files after mutations.
> /plugin marketplace add JeongJaeSoon/agent-guard> /plugin install agent-guard@agent-guard
Repo: JeongJaeSoon/agent-guard
What's inside
Agent Guard is a local-first guardrail for Claude Code, Codex, Git hooks,
GitHub Actions, and direct shell use. It blocks common secret-exposure paths
before a supported tool runs, redacts supported tool output, and scans changed
files after mutations. It uses gitleaks and portable shell; it has no hosted
account, telemetry collector, or developer service.
It is a defense-in-depth boundary. Keep GitHub Secret Scanning, Push Protection, review, and normal credential management in place. Agent Guard is not a vault, DLP system, EDR, or credential rotator.

| I need to… | Start here |
|---|---|
| Install a Claude Code or Codex plugin, CLI, Git hook, or Action | Installation |
| Understand host coverage and configure an integration | Integrations |
| Know which tool routes output masking reaches and how hosts differ | Output masking coverage |
| Verify a setup and understand what the result proves | Verification |
| Configure policy, PII, redaction, or infrastructure behavior | Configuration |
| Deploy for a managed team or troubleshoot an environment | Operations; the Korean/Japanese deployment guide is a supplementary, currently published walkthrough |
After installing a plugin, run its guided setup:
# Claude Code
/agent-guard:setup-agent-guard
# Codex
$setup-agent-guard
The setup flow checks local dependencies and runs synthetic checks. It does not prove that your current host tool route dispatches hooks; complete the harmless live probes in Verification before relying on a plugin boundary.
For a direct repository check:
agent-guard scan-working-tree
agent-guard scan-staged
agent-guard scan-path .
Save the metadata-only local support log with one command:
agent-guard logs export --output agent-guard-support.jsonl
The parent directory must already exist. Agent Guard creates a new mode-0600
file and refuses to replace an existing file or symlink. Use the plugin-local
executable printed by the setup skill when a plugin-only installation does not
provide agent-guard on PATH. See Support for what to submit.
Supported platforms are macOS and Linux on x64 and arm64. Runtime dependencies
are sh, awk, git, jq, and gitleaks 8.30 or newer. Windows is not
currently supported. One of setsid (from util-linux on Linux) or Perl is
also required to run the gitleaks version probe in its own process group.
agent-guard doctor reports a missing isolation tool with the install command
for your platform.
Default processing is local and ephemeral. Read Privacy before enabling an endpoint-backed PII provider. Read Security for responsible disclosure and Support for safe reports.
FAQ
agent-guard is a Claude Code plugin with 2 hand-picked skills for security work, indexed on Flowy. Install it with the command on its page. It includes setup-agent-guard, setup-shell. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
Is this plugin yours?
Claim it with GitHubSubmit a pluginPromote it