Agent Plugins Marketplace
← All plugins

reflex

v0.14.0

Pre-execution risk gate and prompt injection guard for Codex CLI: judges every Bash command and subagent spawn before it runs, with local rules, an editable policy, optional TypeSafe Jev (System One) judgments, System 2 escalation and a human approval queue. Shadow mode by default, no API key needed.

CodexClaude Code1 Skill1 MCP serverstdio

By Ciprian UrsuLicense: MIT1 GitHub starsUpdated 2 hours ago

Directory evidence

Runtimes
Codex and Claude Code
Parsed components
2 skill or MCP entries
Source updated
Sep 29, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install reflex for Codex and Claude Code

Installs for the current user
codex plugin marketplace add ursuciprian/reflex
codex plugin marketplace upgrade reflex
codex plugin add reflex@reflex

Paste and run these commands in a terminal with Codex. They add and refresh the reflex catalog, then install this plugin.

Compatibility: the page URL and API slug “reflex” remain stable.

  • Codex: reflex@agent-plugin-marketplace → reflex@reflex

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/ursuciprian/reflex

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The repository root is the plugin root.

Plugin files

reflex/
├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── skills/reflex/SKILL.md
└── .mcp.json

Included Skills1

reflexskills/reflex/SKILL.md

How to work with the Reflex pre-execution gate in Claude Code or Codex CLI. Use when a shell command was denied or asked by Reflex, when unsure whether a shell command is risky before running it, or when the user asks what Reflex would have blocked in past sessions.

MCP servers1

reflexstdio
command
node
args
${CLAUDE_PLUGIN_ROOT}/mcp.mjs

MCP configuration uses runtime-provided plugin path placeholders such as ${PLUGIN_ROOT} or ${CLAUDE_PLUGIN_ROOT}. Review the manifest for the runtime-specific expansion rules.

Plugin manifests2

.codex-plugin/plugin.json
{
  "name": "reflex",
  "version": "0.14.0",
  "description": "Pre-execution risk gate and prompt injection guard for Codex CLI: judges every Bash command and subagent spawn before it runs, with local rules, an editable policy, optional TypeSafe Jev (System One) judgments, System 2 escalation and a human approval queue. Shadow mode by default, no API key needed.",
  "author": {
    "name": "Ciprian Ursu",
    "url": "https://github.com/ursuciprian"
  },
  "homepage": "https://github.com/ursuciprian/reflex#readme",
  "repository": "https://github.com/ursuciprian/reflex",
  "license": "MIT",
  "keywords": [
    "codex",
    "codex-cli",
    "codex-plugin",
    "codex-hooks",
    "codex-guardrails",
    "pre-tool-use",
    "pre-execution-hook",
    "command-approval",
    "guardrails",
    "agent-security",
    "prompt-injection",
    "prompt-injection-protection",
    "human-in-the-loop",
    "tool-call-gating",
    "devsecops",
    "jev",
    "typesafe"
  ],
  "hooks": "./hooks/codex.json",
  "interface": {
    "displayName": "Reflex",
    "shortDescription": "Pre-execution risk gate and prompt injection guard",
    "longDescription": "Judges every Bash command and subagent spawn before it runs, and screens tool results and prompts for prompt injection. Local rules, shadow mode by default, no API key needed.",
    "developerName": "Ciprian Ursu",
    "category": "Security",
    "websiteURL": "https://github.com/ursuciprian/reflex"
  }
}
.claude-plugin/plugin.json
{
  "name": "reflex",
  "displayName": "Reflex",
  "version": "0.14.0",
  "description": "Pre-execution risk gate and prompt injection guard for Claude Code: judges every Bash command and subagent spawn before it runs, with local rules, an editable policy, optional TypeSafe Jev (System One) judgments, System 2 escalation and a human approval queue. Shadow mode by default, no API key needed.",
  "author": {
    "name": "Ciprian Ursu",
    "url": "https://github.com/ursuciprian"
  },
  "homepage": "https://github.com/ursuciprian/reflex#readme",
  "repository": "https://github.com/ursuciprian/reflex",
  "license": "MIT",
  "keywords": [
    "claude-code",
    "claude-code-hooks",
    "claude-code-plugin",
    "claude-code-security",
    "pre-tool-use",
    "pre-execution-hook",
    "command-approval",
    "guardrails",
    "agent-security",
    "prompt-injection",
    "prompt-injection-protection",
    "human-in-the-loop",
    "tool-call-gating",
    "permission-prompts",
    "devsecops",
    "jev",
    "typesafe"
  ]
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[reflex on Agent Plugins Marketplace](https://pluginsmp.com/plugins/reflex)