Browse all
Edict Compiler for Agentic WASM Code Generation
This compiler pipeline processes JSON ASTs generated by AI agents through multiple stages—including structural validation, type checking, effect analysis, and Z3 contract verification—before generating secure WASM bytecode. It is designed f…
Persistent memory and context management for agents
This comprehensive system provides persistent, structured memory management for agents, enabling the logging of rules, decisions, and mistakes. It includes robust session mining capabilities to analyze historical conversations and predict n…
Turborepo Monorepo Build System Guidance
Provides expert guidance for configuring and managing Turborepo monorepos, including task pipelines, caching strategies, and dependency graphs. It covers best practices for package-specific tasks, CI optimisation, and efficient filtering us…
Vercel React and Next.js Performance Guide
This skill provides comprehensive, Vercel-maintained guidelines for optimizing React and Next.js applications. It covers critical patterns across server-side, client-side, and rendering performance to ensure optimal bundle size and load tim…
React Component Architecture and Composition Patterns
This skill provides architectural guidelines for building scalable and maintainable React components. It advises developers on advanced patterns, such as using compound components and lifting state, to avoid boolean prop proliferation and i…
Guide for Creating Specialized Agentic Skills
This skill provides structured guidance for developing specialized agentic skills, allowing agents to extend core capabilities with domain expertise, complex workflows, and tool integrations. It details best practices for context management…
Interactive pnpm dependency upgrade skill
This skill facilitates interactive dependency updates within a pnpm workspace, managing version bumps and checking against minimum release age constraints. It handles transitive dependency resolution and ensures lockfile integrity without m…
Deduplicate and report bugs to Linear
This skill manages the process of triaging measured bug or regression evidence by searching Linear for existing issues. It requires explicit human approval before either commenting on related tickets or creating new, evidence-only bug repor…
Detect Production Regressions in Langfuse
Proactively identifies production regressions by comparing recent Datadog error logs, spans, and API latency against baseline benchmarks across multiple environments. It generates a structured findings table for human review before optional…
Browser Review for Frontend UI Changes
This skill provides a structured workflow for reviewing user-visible frontend changes, ensuring layout, styling, and navigation regressions are caught. It utilizes the Playwright MCP server to facilitate final signoff for UI-affecting work.
datadog-powered production issue debugging skill
This skill facilitates the root-cause analysis of production failures by integrating Datadog's telemetry (APM, logs, metrics) with the Langfuse codebase. It processes incident reports or issue IDs to deliver a structured analysis, including…
Langfuse Datadog Production Query Recipes
Provides predefined Datadog query shapes for investigating Langfuse production telemetry across multiple environments. It facilitates research into tenant activity, API usage, queue behaviour, and system metrics.
Comprehensive Code Review Workflow for Changes
This skill guides the comprehensive review of code changes (PRs, diffs) focusing on correctness, behavioral regressions, security risks, and performance. It enforces structured output, requiring findings to be listed by severity with precis…
Langfuse Backend Development Guidelines
Comprehensive development guidelines for the Langfuse monorepo, covering tRPC routers, BullMQ processors, and database access patterns using Prisma and ClickHouse. It provides standards for architecture, middleware, and testing across web, …
ClickHouse Schema, Query, and Data Best Practices Review
This skill provides comprehensive, rule-based guidance for optimizing ClickHouse database interactions. It enforces best practices across schema design, query writing, and data ingestion, ensuring all recommendations are cited against speci…
Generate user-focused changelog entries
This skill guides the drafting of user-facing changelog entries for completed features. It requires analyzing a feature branch diff and existing documentation patterns to produce structured, user-centric content.
Analyze Langfuse Cloud Infrastructure Costs
This skill provides evidence-backed cost analysis of Langfuse Cloud infrastructure by querying Metabase cost marts. It identifies total spend, provider splits (e.g., AWS vs ClickHouse), and top cost drivers across various services and usage…
Agent Setup and Maintenance Workflow
A workflow for managing shared agent configurations, skills, and discovery surfaces within the Langfuse repository. It provides guidelines for maintaining canonical files, updating sync scripts, and ensuring robust agent setup during instal…
Managing LLM Model Pricing Configurations
This skill guides developers through updating model pricing configurations across multiple providers, including OpenAI, Anthropic, and Gemini. It covers editing pricing JSON, shared LLM types, and complex regex match patterns to ensure accu…
NocturnusAI Knowledge Base Management
Manage a logical knowledge base by asserting facts, defining Horn clause rules, and performing inference-based queries. The interface supports bulk operations, pattern-based retraction, and schema discovery via an MCP-compatible interface.
Advanced Logical Reasoning for Knowledge Graphs
This tool facilitates advanced logical inference by implementing Negation-as-Failure, managing hypothetical knowledge branches through scoped operations, and supporting probabilistic reasoning via confidence scoring.
Securely fetch and clean web content for LLMs
This utility fetches URLs, providing clean, markdown-formatted content alongside structured metadata and external links. It includes advanced injection safety scanning and handles common web obstacles like paywalls and bot blocks.
Advanced Agent Memory and Context Management
This tool provides comprehensive memory lifecycle management for AI agents, enabling the assertion of facts with temporal metadata and salience scoring. It supports context window population, point-in-time temporal queries, and automated de…
Agentic AI memory and reasoning backend
This tool provides a deterministic backend for agentic AI, managing semantic memory, truth maintenance, and multi-step reasoning. It exposes core functionality via MCP, REST endpoints, and dedicated SDKs for fact assertion, querying, and ru…