Loading...
Loading...
Found 41 Skills
LLM guardrails with NeMo, Guardrails AI, and OpenAI. Input/output rails, hallucination prevention, fact-checking, toxicity detection, red-teaming patterns. Use when building LLM guardrails, safety checks, or red-team workflows.
Search the web for information using DuckDuckGo or other search engines via curl. Use when users ask questions requiring up-to-date information, research, or fact-checking.
Three-layer verification pipeline for AI output. Extracts verifiable claims, finds supporting or contradicting sources via web search, runs adversarial review for hallucination patterns, and produces a structured verification report with source links for human review.
Verify claims in generated output against sources. Use as a separate pass AFTER content generation to catch hallucinations. Critical constraint - cannot be reliably combined with generation in a single pass.
Extracts verifiable claims from ALL .md files (paths, versions, counts, configs, names, endpoints), verifies each against codebase, cross-checks between documents for contradictions.
Verifies factual claims in documents using web search and official sources, then proposes corrections with user confirmation. Use when the user asks to fact-check, verify information, validate claims, check accuracy, or update outdated information in documents. Supports AI model specs, technical documentation, statistics, and general factual statements.
Research topics with web search. Use when: researching a topic or concept, finding current information, answering factual questions, comparing options or technologies. Triggers: research [topic], find out about, what are the best practices for, research the latest on.
Extract falsifiable ideas from input, deep-research each one, and return evidence for or against with strength ratings. Use when user says "find evidence for this", "is this true?", "back this up with data", or "fact-check these claims". Honest about when evidence contradicts the idea.
Fact-checks LLM responses by extracting verifiable claims, verifying each via web search, producing an audit report with verdicts, and optionally revising inaccurate responses. Use when the user asks to audit, fact-check, double-check, or verify a response.
Detects fabricated content, false citations, and unverifiable claims in agent outputs. Uses source verification and consistency checking. Activate on 'detect hallucination', 'fact check', 'verify claims', 'check accuracy', 'find fabrications'. NOT for validation (use dag-output-validator) or confidence scoring (use dag-confidence-scorer).
News Briefing + Verification Workflow. This workflow applies when users request news, headlines, daily briefings, "today's news", "latest updates", breaking news, or specific current figures/events that demand reliable sources and timestamps. Source links and local publication time are mandatory; NEVER fabricate any content.
Use when delegating research tasks requiring verified information from multiple authoritative sources - crawls and fact-checks beyond surface-level findings, providing evidence of verification process with confidence levels for each claim