Prompt Guard
Screens prompts and third-party data for injection and jailbreak attempts before they reach an LLM
- Repository
- Orchestra-Research/AI-Research-SKILLs
- Stars
- 14,195
- Licence
- MIT
- Files
- 1 file
What it does
This skill teaches the agent how to use Meta's Prompt Guard model, a small 86M-parameter classifier, to check text for hidden instructions and jailbreak attempts. The agent can write code that scores user messages, API responses, or retrieved documents and blocks or flags those that look like attacks before they reach a language model. It does not change how the agent itself behaves; it helps developers build a safety filter into their own LLM application.
Great for
- Developers building LLM apps that accept user input or pull in web, API, or RAG documents
- Teams wanting a lightweight, CPU-friendly first-pass filter for prompt injection
- Builders who need multilingual screening across the eight supported languages
Skip it if
- Content moderation of violence, hate, or criminal planning, where the skill itself points to LlamaGuard instead (wrong tool for that job)
- Production security on its own: the skill reports 3.9% false positives out of distribution and misses are possible, so it should be one layer among several
- Non-developers wanting a ready-made app; this is code guidance, not a finished product
Try asking
- “Add a jailbreak filter in front of my chatbot's LLM call using Prompt Guard”
- “Screen retrieved RAG documents in batches for embedded instructions before summarising them”
- “Check API responses from third-party services for injected commands”
- “Set up a defense-in-depth pipeline combining Prompt Guard with LlamaGuard”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
prompt-guard.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip prompt-guard.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip prompt-guard.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Use the default 0.5 threshold for direct user input and a lower threshold such as 0.3 for third-party data, as the skill suggests, and tune both on your own traffic.
- Treat the classifier as one layer: combine it with content moderation and output validation rather than relying on it alone.
- Test with examples from your own application, since reported accuracy figures come from Meta's benchmarks and may differ in your domain.
What's inside
- SKILL.md
Redistributed unmodified from Orchestra-Research/AI-Research-SKILLs under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.