Prompt Guard
Prompt Guard is an agent skill (a SKILL.md file) from Orchestra-Research/AI-Research-SKILLs. Meta's 86M prompt injection and jailbreak detector. It works with Claude Code, Codex, Cursor, Gemini CLI and OpenCode and has 13,315 GitHub stars across a repository of 5 listed skills.
- Multi-skill repo
- Plugin marketplace
- Security
- Maintained
Add this skill
Claude
This repository is a Claude Code plugin marketplace. In Claude Code:
/plugin marketplace add Orchestra-Research/AI-Research-SKILLs
/plugin # browse ai-research-skills and install the plugin that contains prompt-guardIn the Claude apps, zip the prompt-guard folder and upload it under Customize > Skills > + > Upload a skill (code execution must be on).
ChatGPT / Codex
Codex reads skills from .agents/skills/ in a repo or ~/.agents/skills/ for every project:
git clone --depth 1 https://github.com/Orchestra-Research/AI-Research-SKILLs.git
cp -r AI-Research-SKILLs/07-safety-alignment/prompt-guard .agents/skills/prompt-guard # repo; ~/.agents/skills for all projectsStandalone skills also load in the ChatGPT desktop app.
Cursor
Cursor loads skills from .cursor/skills/ (or ~/.cursor/skills/) and also reads .claude/skills/:
git clone --depth 1 https://github.com/Orchestra-Research/AI-Research-SKILLs.git
cp -r AI-Research-SKILLs/07-safety-alignment/prompt-guard .cursor/skills/prompt-guard # project; ~/.cursor/skills for all projectsSource (checked Oct 7, 2026): code.claude.com/docs/en/skills (opens in a new tab), code.claude.com/docs/en/plugin-marketplaces (opens in a new tab), support.claude.com/en/articles/12512180-using-skills-in-claude (opens in a new tab), learn.chatgpt.com/docs/build-skills (opens in a new tab), cursor.com/docs/context/skills (opens in a new tab)
What this skill does
Meta's 86M prompt injection and jailbreak detector. Filters malicious prompts and third-party data for LLM apps. 99%+ TPR, <1% FPR. Fast (<2ms GPU). Multilingual (8 languages). Deploy with HuggingFace or batch processing for RAG security. Prompt Guard is an 86M parameter classifier that detects prompt injections and jailbreak attempts in LLM applications.
More skills in Orchestra-Research/AI-Research-SKILLs
5 skills are listed from this repository.
Similar skills
More security skills
Wizard
mattpocock/skills
Generate an interactive bash wizard that walks a human through steps only they can perform.
SecurityShellPostgres Patterns
affaan-m/ECC
PostgreSQL database patterns for query optimization, schema design, indexing, and security.
SecurityJavaScriptPonytail Audit
DietrichGebert/ponytail
Whole-repo audit for over-engineering. Like ponytail-review, but scans the entire codebase instead of a diff: a ranked list of what to delete, simplify, or replace with stdlib/native equivalents.
SecurityJavaScriptNext Bundle Optimizer
vercel/next.js
Audit and reduce Next.js browser initial-load work.
OfficialSecurityJavaScriptCloud
browser-use/browser-use
Documentation reference for using Browser Use Cloud — the hosted API and SDK for browser automation.
OfficialSecurityPythonSecurity And Hardening
addyosmani/agent-skills
Hardens code against vulnerabilities. Use when auditing an input handler for vulnerabilities, when handling user input, authentication, data storage, or external integrations, or when checking a…
SecurityJavaScript
What is the Prompt Guard skill?
Prompt Guard is an agent skill (a SKILL.md file) from Orchestra-Research/AI-Research-SKILLs. Meta's 86M prompt injection and jailbreak detector. It works with Claude Code, Codex, Cursor, Gemini CLI and OpenCode and has 13,315 GitHub stars across a repository of 5 listed skills. Its SKILL.md lives at github.com/Orchestra-Research/AI-Research-SKILLs/07-safety-alignment/prompt-guard.
How do I install the Prompt Guard skill?
In Claude Code, run /plugin marketplace add Orchestra-Research/AI-Research-SKILLs, then install its plugin from /plugin. For Codex or Cursor, copy the prompt-guard folder into .agents/skills/ or .cursor/skills/.
Is the Prompt Guard skill free?
Yes. The repository is open source under the MIT license.
Is Prompt Guard maintained?
The repository's most recent commit was on Jun 16, 2026. Its latest release is v1.7.2. appsgit only lists skills from repositories with a commit in the last six months.