Skills tagged "ai-security"

red-teaming-llms-with-garak - Red-Teaming LLMs with garak

ai-security llm-red-teaming garak prompt-injection

★ 2 · Updated 2026-09-24

Runs NVIDIA garak probe suites against an LLM endpoint and interprets the resulting hit-rate report.

⚙ Enumerate probes and detectors ⚙ Configure target LLM endpoints ⚙ Run security probe suites

assessing-vector-and-embedding-weaknesses - Assess security weaknesses in vector stores and embeddings

ai-security vector-database embedding-inversion rag-security

★ 3 · Updated 2026-09-24

Test vector stores for embedding inversion, cross-tenant leakage, membership inference, and knowledge-base poisoning.

⚙ Measure embedding inversion exposure ⚙ Run membership inference probes ⚙ Test multi-tenant isolation controls

llm-security - LLM and AI Agent Security Assessment

llm-security ai-security prompt-injection agent-security

★ 1 · Updated 2026-09-16

Conduct authorized security assessments of LLM applications and AI agents, covering prompt injection, tool abuse, RAG exposure, and supply-chain risks.

⚙ Reconnoiter AI attack surfaces ⚙ Test prompt injection vulnerabilities ⚙ Test tool abuse risks

mcp-tool-execution - Deterministic MCP Tool Execution

mcp tool-execution parameter-validation deterministic-execution

★ 650 · Updated 2026-05-11

Registers, validates, and executes MCP tools deterministically, enforcing tool-mediated AI interactions.

⚙ register MCP tools ⚙ validate tool definitions ⚙ validate parameters