security-ai-llm
Skill0 starsChanged 16 days agoMITView on GitHub
Review AI/LLM applications for prompt injection, unsafe tool execution, retrieval and tenant-isolation failures, sensitive-data leakage, unsafe output handling, and model-cost abuse. Use when a repository integrates models, agents, RAG, embeddings, or AI APIs.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- lgzarturo/codeconductortheir own version, first seen Oct 18, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- lgzarturo/codeconductor★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- ai-security-auditSkill
- security-apiSkill
- security-reconSkill
- security-secretsSkill
- security-supply-chainSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Test by static reasoning and safe local fixtures whether untrusted content can override instructions, invoke unauthorized tools, cross …
- Distinctive: line 18
Create candidates with exact code evidence and impact, then use a separate validation pass to try to disprove each candidate.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AI and LLM Security Audit
Audit the application around the model, not just the prompt text. Trace how untrusted content influences tools, data access, and downstream actions.
Workflow
- Identify model providers, system/developer prompts, user prompt construction, retrieval pipelines, vector stores, agent loops, tool/function definitions, output consumers, logging, and fallback paths.
- Map trust boundaries between users, retrieved documents, external content, the model, tools, application services, and tenant data.
- Test by static reasoning and safe local fixtures whether untrusted content can override instructions, invoke unauthorized tools, cross tenant boundaries, expose secrets or personal data, or cause unsafe downstream actions.
- Inspect tool authorization at execution time. A prompt saying “do not do this” is not a security control; enforce identity, scope, argument validation, and least privilege in code.
- Review retrieval filtering, document provenance, embedding/vector access controls, output validation, HTML/SQL/shell sinks, secret-bearing logs, data retention, and resource/cost limits.
- Distinguish model uncertainty from code-level vulnerabilities. Demonstrate the actual path from attacker-controlled content to a sensitive operation.
- Create candidates with exact code evidence and impact, then use a separate validation pass to try to disprove each candidate. Never report a hypothetical model response as a confirmed exploit without a reproducible path.
- Redact secrets and private data. Use only synthetic fixtures for prompt-injection or data-exfiltration demonstrations.
References
- Core workflow:
../../SKILL.md - AI playbook:
../../agents/ai-llm-security.md - Injection playbook:
../../agents/injection.md - AI rules:
../../rules/ai.md - Validator protocol:
../../validation/finding-validator.md
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.