LLM guardrails for security — input/output validation with NeMo Guardrails, Presidio, and Guardrails AI
Harden LLM apps with input/output guardrails — NeMo Guardrails (Colang) flows, Python PII validators, and Guardrails AI schema checks against prompt injection, PII leaks, and hallucinations
- What
- Harden LLM apps with input/output guardrails — NeMo Guardrails (Colang) flows, Python PII validators, and Guardrails AI schema checks against prompt injection, PII leaks, and hallucinations
- Cost
- Free
- Needs
- a Python 3.10+ environment with pip; an OpenAI API key (billed) OR a local LLM endpoint for the NeMo Guardrails self-check rails
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor — @mukul975's guardrails skill: a complete defensive playbook for adding input/output safety controls to LLM applications. The agent installs the guardrail frameworks (NeMo Guardrails for Colang-based rail flows, Guardrails AI for structured output validation, Microsoft Presidio and spaCy for PII detection), runs a guardrails pipeline agent in multiple modes (full, input-only, output-only, PII redaction, JSON for dashboards), authors JSON content policies (allowed/blocked topics, blocked patterns, PII categories), writes Colang 2.0 flow definitions (input rails like self-check, jailbreak check, sensitive-data masking; output rails like hallucination check), and deploys the whole thing as validation middleware around an LLM app or RAG pipeline. It ships key-concept explanations, a tools inventory, verification checks (injection patterns blocked, PII redacted, <200 ms input-only latency), and an honest scope boundary: guardrails are defense-in-depth, not a replacement for authentication, authorization, or network security. Honest caveats: needs Python 3.10+ and real installs (nemoguardrails, presidio, a spaCy model); the self-check rails call a model — an OpenAI API key (billed) or a local LLM endpoint, so costs are avoidable with a local model. Apache-2.0 licensed (frontmatter and manifest agree). Skill Harbor never reviews the code, review it yourself before use. Discovered via skills.sh.
Version:
Quick answer
LLM guardrails for security is a free Muse skill for hardening LLM apps with input and output guardrails. It covers NeMo Guardrails flows in Colang, Python PII validators and Guardrails AI schema checks. Those layers defend against prompt injection, PII leaks and hallucinations, keeping LLM applications safer in production use.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit 54a7988Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: a Python 3.10+ environment with pip; an OpenAI API key (billed) OR a local LLM endpoint for the NeMo Guardrails self-check rails Install "LLM guardrails for security — input/output validation with NeMo Guardrails, Presidio, and Guardrails AI" for me. It gives my agent @mukul975's defensive guardrails playbook: install the guardrail frameworks (nemoguardrails, guardrails-ai, presidio-analyzer, presidio-anonymizer, spacy + a language model), run the guardrails security agent in full/input-only/output-only/PII/JSON modes, author a JSON content policy (allowed and blocked topics, blocked patterns, PII categories, output limits), define Colang 2.0 rail flows (self check input, check jailbreak, mask sensitive data, self check output, hallucination check), integrate the pipeline as validation middleware around an LLM app or RAG pipeline, and review guardrail logs for block rates and bypass attempts. Apache-2.0 licensed. Repository: https://raw.githubusercontent.com/mukul975/anthropic-cybersecurity-skills/54a798831d2266a3ca61ce68a7acb80b81160d57/skills/implementing-llm-guardrails-for-security/SKILL.md Pinned version: install the code at commit 54a798831d2266a3ca61ce68a7acb80b81160d57 exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch the SKILL.md file (and any helper files) from the repository path into a temporary folder 2. Safety check: review the SKILL.md and scripts for anything suspicious (unexpected network calls, shell commands, credential harvesting). This repo should contain zero secrets in code, credentials only via environment variables, allowed hosts declared in the SKILL.md. Verify that holds here; STOP on any red flag and tell me. 3. Install it as a skill: copy SKILL.md and its helper files into the agent's skills directory, in a folder named "implementing-llm-guardrails-for-security". 4. Verify with no network calls: frontmatter valid, files in place. 5. Report what was installed, where, and what I still need to do myself (e.g. pip install the guardrail packages and a spaCy model, set OPENAI_API_KEY or point the self-check rails at a local LLM endpoint, author my content policy JSON). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
- It works with an API key or a token that you provide.
What the scan found — excerpts from the install prompt
- uses API keys or tokens — Prerequisites: a Python 3.10+ environment with pip; an OpenAI API key (billed) OR a local LLM endpoint for the · Prérequis : un environnement Python 3.10+ avec pip ; une clé API OpenAI (facturée) OU un endpoint LLM local po
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- 54a7988 · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : uses API keys or tokens
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.