LLM, RAG, and Agent Security: Prompt Injection, Tool Permissions, and Boundary-Aware Defense
AI Learning Project
LLM, RAG, and Agent Security: Prompt Injection, Tool Permissions, and Boundary-Aware Defense
A practical guide to LLM/RAG/Agent security covering prompt injection, external content isolation, tool allowlists, approval gates, and boundary-aware defenses.
AI Learning Project

LLM, RAG, and Agent Security: Prompt Injection, Tool Permissions, and Boundary-Aware Defense

Reading info

Level: Professional Reading time: 12 min
  • LLM Security
  • RAG
  • Agent Tools
  • Prompt Injection
Open knowledge map

The security boundary of an LLM application does not end at the model. RAG documents, system prompts, tool calls, agent memory, external plugins, and human approvals all influence behavior. Prompt injection risk appears when untrusted text is treated as high-priority instruction.

This article explains prompt injection, tool permissions, and boundary-aware controls for RAG and agent systems. The lab uses a deterministic toy simulator. It does not call a real LLM, access the network, or include real exploit payloads.

1. Instructions and data must be separated

A secure LLM application should distinguish three content classes:

  • System instruction: developer-defined behavior boundaries and safety policy.
  • User request: the task the user wants completed.
  • External data: RAG retrieval, web pages, email, PDFs, and database results.

Prompt injection occurs when external data tries to promote itself into an instruction. For example, a retrieved document may try to override rules and trigger an administrative tool. The control objective is not to filter every possible phrase. It is to ensure external data never grants tool authority.

2. RAG risk path

A typical RAG risk chain is:

untrusted document
  -> retriever
  -> prompt context
  -> model reasoning
  -> tool call or sensitive answer

If the chain has no boundary controls, instructions inside untrusted documents can be mixed with trusted system policy.

3. Local guard experiment

The lab script contains three toy documents. One comes from an untrusted source and contains an inert injection-like string. The guard blocks it by source and risk pattern.

cd ai-security-lab
python src/rag_prompt_injection_guard_demo.py --quick --out results/rag-guard-results.csv

The output contains guard_enabled, blocked_documents, unauthorized_tool_call_attempt, and answer. The demo does not model real LLM capability. It demonstrates an engineering boundary: untrusted data must not directly authorize high-risk tools.

4. Agent tool permission design

Agent systems must separate “the model wants to call a tool” from “the system permits the tool call”. Recommended rules:

  • Classify tools by risk: read-only, write, external send, delete, payment, and permission change.
  • Use allowlists by default; do not let the model freely discover high-risk tools.
  • Require structured parameter validation and human approval for high-risk actions.
  • Mark tool results as data before returning them to the model; they are not new system instructions.
  • Log every tool call with source context and authorization reason.

5. Engineering controls

  • Record document source, trust level, and update time in retrieval.
  • Mark external content as data, not instruction, inside prompt templates.
  • Use source filtering, pattern scans, and summarization isolation for untrusted sources.
  • Perform server-side policy checks for tools instead of relying on model self-restraint.
  • Maintain a prompt-injection regression set from observed failures.

6. Limitations

String scanning cannot cover every indirect prompt injection. Real systems also need model behavior evaluation, tool sandboxing, permission audits, minimized data returns, and human approval. This demo shows a minimum boundary design, not a complete solution.

7. References

Internal links

Related tutorials

Review foundation LLM Visualizer

Interactively inspect tokenization, embeddings, attention, next-token sampling, and KV cache without loading a real model.

 Learning path AI Learning Path

Read AI foundations, machine learning, CNN work, traffic defense, and AI security in order.

 Resources Resource Library

Download companion code, datasets, diagrams, model files, and notes.

 Topic hub Student AI Projects

Connect handwritten digits, browser demos, and CNN projects into a portfolio path.

 Knowledge map Knowledge Map

See how this article connects to projects, resources, and learning routes.

Search questions

FAQ

Who is this article for?

This article is for readers who want a professional-level guide to LLM, RAG, and Agent Security. It takes about 12 min and focuses on LLM Security, RAG, Agent Tools, Prompt Injection.

What should I read next?

Use the related tutorials and project links below the article to continue through the closest topic hub.

Does this article include runnable code or companion resources?

Yes. Use the run notes, resource cards, and download links on the page to reproduce the example or inspect the companion files.

How does this article fit into the larger site?

It is connected to the article context block, learning routes, resources, and project timeline so readers can move from concept to implementation.

Article context

AI Learning Project

A practical route from AI concepts to machine learning workflow, evaluation, neural networks, Python practice, handwritten digits, a CIFAR-10 CNN, adversarial traffic-defense notes, and AI security.

Level: Professional Reading time: 12 min
  • LLM Security
  • RAG
  • Agent Tools
  • Prompt Injection

Your next step

Open resources
Open resource View project
Other language version LLM/RAG/Agent 安全：Prompt Injection、工具权限和边界感知防护
Share summary LLM, RAG, and Agent Security

Separate instructions from data and enforce tool permissions against indirect prompt injection.

Download share card Open share center

Companion resources

Leave a Reply

Project timeline

Published posts

  1. AI Basics Learning Roadmap Separate AI, machine learning, and deep learning before going into implementation details.
  2. Machine Learning Workflow Follow the practical path from data and features to training, prediction, and evaluation.
  3. Model Training and Evaluation Understand loss, overfitting, train/test splits, accuracy, recall, and F1.
  4. Neural Network Basics Move from perceptrons to activation, forward propagation, backpropagation, and training loops.
  5. NLP Basics: Understanding Bag of Words and TF-IDF An introduction to the most fundamental text representation methods in NLP: Bag of Words (BoW) and TF-IDF.
  6. RNN Basics: Handling Sequential Data with Memory Understand the core concepts of Recurrent Neural Networks (RNN), the role of hidden states, and their application in NLP.
  7. Transformer Self-Attention Read Q/K/V, scaled dot-product attention, multi-head attention, and positional encoding before exploring LLM internals.
  8. Python AI Mini Practice Run a small scikit-learn classification task and read the experiment output.
  9. Handwritten Digit Dataset Basics Read train.csv, test.csv, labels, and the flattened 28 by 28 pixel layout before training the classifier.
  10. Handwritten Digit Softmax in C Follow the C implementation from logits and softmax probabilities to confusion matrices and submission export.
  11. Handwritten Digit Playground Notes See how the offline classifier was adapted into a browser demo with drawing input and probability output.
  12. CIFAR-10 Tiny CNN Tutorial in C Build and train a small convolutional neural network for CIFAR-10 image classification, then read its loss and accuracy output.
  13. Building a Tiny CIFAR-10 CNN in C: Convolution, Pooling, and Backpropagation A source-based walkthrough of cifar10_tiny_cnn.c, covering CIFAR-10 binary input, 3x3 convolution, ReLU, max pooling, fully connected logits, softmax, backpropagation, and local commands.
  14. High-Entropy Traffic Defense Notes Study encrypted metadata leaks, entropy, traffic classifiers, and a defensive Python chaffing prototype.
  15. AI Security Threat Modeling Build a defense map with NIST adversarial ML, MITRE ATLAS, and OWASP LLM risks.
  16. Adversarial Examples and Robust Evaluation Evaluate clean and perturbed accuracy with an FGSM-style digits experiment.
  17. Data Poisoning and Backdoor Defense Study poison rate, trigger behavior, attack success rate, and training pipeline controls.
  18. Model Privacy and Extraction Defense Measure membership inference signal and surrogate fidelity against a local toy model.
  19. LLM, RAG, and Agent Security Separate instructions from data and enforce tool permissions against indirect prompt injection.

Published resources

  1. Python AI practice code guide The article includes a runnable scikit-learn classification script.
  2. digit_softmax_classifier.c The C source for the handwritten digit softmax classifier.
  3. train.csv.zip Compressed handwritten digit training set with 42000 labeled samples.
  4. test.csv.zip Compressed handwritten digit test set with 28000 unlabeled samples.
  5. sample_submission.csv The official submission format example for checking the final output columns.
  6. submission.csv The prediction file generated by the current C project.
  7. digit-playground-model.json The compact softmax demo model and sample set used by the browser playground.
  8. digit-sample-grid.svg A small handwritten digit preview grid extracted from the training set.
  9. Handwritten digit project bundle Contains the source file, compressed datasets, submission files, browser model, and preview grid.
  10. cifar10_tiny_cnn.c source Single-file C tiny CNN with CIFAR-10 loading, convolution, pooling, softmax, and backpropagation.
  11. model_weights.bin sample weights Model weights generated by one local small-sample run.
  12. test_predictions.csv sample predictions Sample test prediction output from the CIFAR-10 tiny CNN.
  13. CNN project explanation PDF Companion explanation material for the CNN project.
  14. Virtual Mirror redacted code skeleton A redacted mld_chaffing_v2.py control-flow skeleton with secrets, node topology, and target lists removed.
  15. Virtual Mirror stress-test template A redacted CSV template for CPU, memory, peak threads, pulse rate, latency, and error measurements.
  16. Virtual Mirror classifier-evaluation template A CSV template for TP, FN, FP, TN, accuracy, precision, recall, F1, ROC-AUC, entropy, and JS divergence.
  17. Virtual Mirror resource notes Notes explaining why the public resources include only redacted code, test templates, and architecture context.
  18. AI Security Lab README Setup, safety boundaries, and quick-run commands for the AI Security series.
  19. AI Security Lab full bundle Includes safe toy scripts, result CSVs, risk register, attack-defense matrix, and architecture diagram.
  20. AI security risk register CSV risk register template for AI threat modeling and release review.
  21. AI attack-defense matrix Maps attack surface, toy demo, metric, and defensive control into one CSV table.
  22. AI Security Lab architecture diagram Shows threat modeling, robustness, data integrity, model privacy, and RAG guardrails.
  23. FGSM digits robustness script FGSM-style perturbation and accuracy-drop experiment for a local digits classifier.
  24. Data poisoning and backdoor toy script Demonstrates poison rate, trigger behavior, and attack success rate on digits.
  25. Model privacy and extraction toy script Outputs membership AUC, target accuracy, surrogate fidelity, and surrogate accuracy.
  26. RAG prompt injection guard toy script Uses a deterministic toy agent to demonstrate external-data demotion and tool-policy blocking.
  27. Deep Learning topic share card A 1200x630 SVG card for sharing the Deep Learning / CNN topic hub.
  28. Machine Learning From Scratch share card A 1200x630 SVG card for the K-means, Iris, and ML workflow topic hub.
  29. Student AI Projects share card A 1200x630 SVG card for handwritten digits, C classifiers, and browser demos.
  30. CNN convolution scan animation An 8-second Remotion animation showing how a 3x3 convolution kernel scans an input and builds a feature map.

Current route

  1. AI Basics Learning Roadmap Learning path step
  2. Machine Learning Workflow Learning path step
  3. Model Training and Evaluation Learning path step
  4. Neural Network Basics Learning path step
  5. Transformer Self-Attention Learning path step
  6. LLM Visualizer Learning path step
  7. Python AI Mini Practice Learning path step
  8. Handwritten Digit Dataset Basics Learning path step
  9. Handwritten Digit Softmax in C Learning path step
  10. Handwritten Digit Playground Notes Learning path step
  11. CIFAR-10 Tiny CNN Tutorial in C Learning path step
  12. High-Entropy Traffic Defense Notes Learning path step
  13. AI Security Threat Modeling Learning path step
  14. Adversarial Examples and Robust Evaluation Learning path step
  15. Data Poisoning and Backdoor Defense Learning path step
  16. Model Privacy and Extraction Defense Learning path step
  17. LLM, RAG, and Agent Security Learning path step

Next notes

  1. Add more image-classification and error-analysis cases
  2. Turn common metrics into a quick reference
  3. Add more AI security defense experiment notes