$ loading_
Explore AI failure mode taxonomy to identify risks and improve systems.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "io.github.elyngved/failmodes" MCP server from askskill: Run: claude mcp add 'io-github-elyngved-failmodes' -- npx -y failmodes-mcp
Using the Fail Modes taxonomy, list common AI failure modes for a customer support chatbot. Group them by input understanding, reasoning, output safety, and user experience, and explain the risk of each group.
A categorized list of failure modes with risk notes for team review.
Refer to the Fail Modes taxonomy and design 10 evaluation test cases for a text generation model, covering factual errors, prompt injection, overconfident answers, and context omission. Explain the goal of each test.
A set of test cases covering key failure modes, each with a clear evaluation goal.
Based on the Fail Modes taxonomy, identify relevant failure modes for hallucinated citations in a retrieval-augmented generation system, and propose detection methods, mitigation strategies, and a pre-launch checklist.
An analysis of relevant failure modes plus actionable detection, mitigation, and pre-launch checks.
Connect AI to a knowledge base for semantic retrieval and compliant content management.
Provide AI agents with coding standards, testing, planning, and requirements guidance.
Analyze and manage software architecture graphs for impact, dependencies, and design decisions.
Connect AI apps to a shared knowledge graph for consistent retrieval and reasoning.
Extract architecture, dependencies, and API knowledge from any codebase quickly.
Track structured reasoning with confidence, branches, and revisions for complex problem solving.