MCP App StoreMCP App Store
from claude-mpm-skills
Techniques and patterns to compress multi-turn AI conversations (summaries, RAG, hierarchical summarization) to reduce token costs and preserve key information.
Session Compression provides concrete patterns and code for reducing conversation token usage via summarization, embedding-based retrieval (RAG), hierarchical and rolling summaries, and prompt caching. It includes progressive compression thresholds, sample implementations (LangChain, Anthropic), and production guidance so agents can manage long-running multi-turn sessions without losing important context.
Use this skill when conversations approach model context limits, for long-running chat sessions (support, tutoring, coding), when token costs are high, or when you need to persist multi-session context. Avoid for short (<10 turns) or verbatim-critical workflows (legal/compliance).
Inferred compatibility: Claude-family assistants (Claude 3.5 Sonnet/Haiku), LangChain integrations, and systems using OpenAI embeddings — useful for agents that can call external LLMs or vector stores.
A comprehensive reference skill covering AI session compression techniques including summarization (extractive, abstractive, hierarchical, rolling), RAG-based retrieval, and hybrid approaches. Contains extensive Python code examples using Anthropic/OpenAI/LangChain APIs. No bundled scripts to execute — purely a knowledge/reference SKILL.md. Placeholder API keys throughout are instructional, not real credentials. Well-structured with progressive disclosure and clear use-case guidance.
No security issues found. All API key references are instructional placeholders. No executable scripts bundled. The skill is a thorough reference document on session compression patterns with well-organized sections covering concepts, implementations, production patterns, and tool recommendations. Architecture follows the skill spec with frontmatter and progressive disclosure. Code examples are solid but some are partial (e.g., _remove_redundant_messages has 'pass'). Useful for developers working on long-context AI applications, though somewhat niche audience.
WordPress Block Editor & Full Site Editing (FSE)
Guidance for building block themes and custom Gutenberg blocks using theme.json, HTML templates, and server-rendered blocks for WordPress 6.7+.
Hono JSX (server-side rendering)
Server-side JSX renderer for Hono: async components, Suspense streaming, head hoisting and error boundaries for fast SSR and HTML generation.
Axum (Rust) — Production Web APIs
Patterns and best-practices for building production-ready Rust HTTP APIs with Axum: typed handlers, extractors, middleware, error handling, tracing, graceful sh
Go CLI (Cobra + Viper)
Guides building production-quality Go CLIs using Cobra for commands and Viper for configuration, with patterns for flags, completion, testing and best practices
MPM Orchestration Demo
Canonical demo of the Command → Agent → Skill orchestration pattern: shows preloaded vs dynamic skill invocation, with a code-review example and templates.
Datadog Observability
Guides setup and best practices for Datadog APM, logs, metrics, synthetics and RUM to implement full-stack monitoring, tracing, and cost optimization in product
Database Migration Patterns
Guided, safe patterns for evolving database schemas in production with decision trees, tooling examples, and rollback strategies.
Python asyncio
Guides for writing concurrent I/O-bound Python code with async/await: event loops, tasks, aiohttp, async DB patterns, and FastAPI async endpoints.
Hono (core) - JavaScript Framework Skill
Skill entry for Hono core framework (JavaScript) — metadata missing in source; flagged as failed.
Caveman Compress
Compress natural-language memory files into a compact 'caveman' format to save LLM tokens while preserving code, links, and structure.
Reppo — AgentMind Publisher
Publish creative content to Moltbook and mint posts as on-chain pods on Reppo.ai (Base) to earn $REPPO through human voting.
Seedance Prompt — Short
Build, validate, and compress short-form Seedance 2.0 prompts (30–100 words) using a layered compression engine and @Tag delegation for quad-modal (T2V/I2V/V2V/
Go Data Structures
Authoritative guidance on choosing and using Go built-in and standard-library data structures, with practical best practices for slices, maps, arrays, container
Claude Code (Clawdbot)
Wrapper to run the local Claude Code CLI in headless or interactive (tmux) modes for code analysis, refactors, tests, and structured output.
Anthropic Security Basics
Practical security practices for Anthropic Claude integrations: API key management, input validation, prompt-injection defenses, and output scanning.
Memory Search
Search past coding sessions and observations using natural-language queries to retrieve decisions, notes, and context.
DSPy — Declarative LM Programming
Use DSPy to build declarative, modular LM pipelines, optimize prompts automatically, and assemble reliable RAG/agent systems with structured signatures and opti