Ai rag defense
📚 Agent skills distilled from technical books — AI Engineering, Context Engineering, Designing Data-Intensive Applications, and more. Agent-agnostic, plain Markdown. Give your AI agent a bookshelf.
npx -y skills add ebarti/skills --skill ai-rag-defenseAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Token reduction, retrieval fidelity, and prompt-injection defense layers for the Context Engine. Covers a Summarizer agent for proactive context reduction, micro-context engineering (prompt design inside agents), high-fidelity RAG with source-metadata citations, layered untrusted-context defenses against prompt injection, and grounded-reasoning validation that prevents hallucination. Use this skill when: - Reducing token costs by adding a Summarizer agent - Designing system prompts inside specialist agents (micro-context) - Adding citations and source metadata to RAG outputs - Defending agents against prompt injection / data poisoning - Validating agent reliability across high-fidelity / summarization / grounded test cases - Preventing hallucination when retrieval returns no result
SKILL.md
2.1 KB, as published. Nobody here has run it
AI RAG Defense
Knowledge from "Context Engineering for Multi-Agent Systems" (Chapters 6-7). The reduction, fidelity, and defense layers that turn a working engine into a trustworthy one.
Quick Start
- Check
guidelines.mdto find which files to load - Load only relevant files (each topic has knowledge.md, rules.md, examples.md)
- Apply guidance to your work
Contents
References
| Category | Purpose |
|---|---|
summarizer-agent | Cost-management agent for proactive context reduction |
micro-context-engineering | Prompt design inside agents (macro vs micro) |
high-fidelity-rag | Source-metadata-on-chunks pattern, citations-mandatory Researcher |
input-sanitization | Untrusted context boundary, prompt-injection threat model + smells |
grounded-reasoning | Report-negative-finding pattern, multi-case validation |
Workflows
| Workflow | Purpose |
|---|---|
workflows/add-summarizer.md | Integrate Summarizer agent for proactive context reduction |
workflows/add-citations.md | Upgrade RAG to high-fidelity (source metadata + cite) |
workflows/defend-against-injection.md | Add untrusted-context boundary + run grounded-reasoning validation suite |
Guidelines
See guidelines.md for task-based file selection.