Failure-first spec stress test — inverts sycophancy to surface failure modes (C2/P1).

MCP server for the Fail Modes taxonomy — a knowledge base of AI system failure modes

Local-only OSS contribution command center with 41 deterministic gates against AI-slop failure modes

Defense against LLM Rust failure modes for Claude Code and Codex: 58 categories of bugs that survive rustc, clippy, and cargo test.

Claude Code skill: Figma → pixel-perfect HTML/CSS with a self-verifying fidelity pipeline (131 documented failure modes, guard-enforced). Running this package installs the skill into ~/.claude/skills.

Content Hooks MCP for creators and AI agents. Short-form content mechanics: 10 hook styles with failure modes, script structures for 5 formats, retention engineering, CTA rules, hook + script auditing. Receipts over hype.

ESLint rules that catch the dominant agentic-coding failure modes — empty handlers (Gate 1) and mock-data leaks into production paths (Gate 2). Designed to slot into the Mushi Mushi v2 CI Gate pipeline.

Mock up a plain or TLS-encrypted TCP connection without using any actual network facilities or needing any permissions. This is useful for testing, particularly playing around with different TLS failure modes.

Stops your AI coding agent from repeating the same mistakes. OpenCode plugin that consolidates failure modes, patterns, and preferences from every session into a persistent knowledge base injected into future sessions automatically.

The ALEF Pattern Catalog — 29 named failure modes for agentic AI systems + 3 core architectural doctrines. Machine-readable JSON-LD, browsable at n50.io/patterns. Cross-referenced to OWASP-ASI06-2025 where applicable. Includes typed reader for embedding t

Adversarial benchmark framework for auditing model failure modes and documenting corrections.

Interceptor SDK for AI agents — tamper-evident evidence capture with zero added failure modes

AI agent rule configuration preset for @lapidist/design-lint — enables all rules targeting AI code generation failure modes.

A tiny (183B to 210B) and fast utility to ascend parent directories

Collection of ported legacy language modes for the CodeMirror code editor

A tiny (304B to 489B) utility for check for deep equality

A tiny (130B to 205B) and fast utility to randomize unique IDs of fixed length

Common library for code that's used across the Client, Worker, and/or Workflow

A tiny (195B to 220B) utility to recursively list all (total) files in a directory

Kill-tested guard skills for AI coding agents — quality gates that catch AI failure modes before the agent says done, with deterministic zero-dep CLI/CI enforcement.