Prompt engineering, RAG, context management, and agent memory.
Everything upstream of the model call: prompt engineering patterns, retrieval and RAG pipelines, context window management, and persistent memory across sessions. These artifacts decide what your agent knows when it starts working, which usually matters more than which model you picked. Includes context compaction, knowledge-base wiring, and memory that survives a restart.
Ranked by GitHub stars. Search to find fast, or page through the full list.
Cloud-backed persistent memory powered by Deeplake — read, write, and share memory across Claude Code sessions and agents
MCP server for retrieving context from a Qdrant vector database
Migrates Claude Code personal memory (the per-project memory/ directory) into tool-agnostic reference docs, so other AI CLIs that auto-load AGENTS.md (Codex primarily; the content architecture transfers to Cursor and others) working in the same directory can read the same user profile, collaboration preferences, and methodology instead of being blind to them. Use this whenever the user says things like "migrate my memory", "my memory is locked to Claude Code", "make Codex/Cursor read my profile", "memory should live in docs not one tool", or reports that a second AI tool doesn't know who they are; also use it when memory has grown bloated with content that should be shared across tools or projects. Covers diagnosis, the references/ + CLAUDE.md-inline + AGENTS.md-symlink architecture, multi-agent review, empirical codex verification, and memory cleanup. Inline only — it orchestrates review subagents and runs codex.
Transform vague prompts into precise, well-structured specifications using EARS (Easy Approach to Requirements Syntax) methodology. This skill should be used when users provide loose requirements, ambiguous feature descriptions, or need to enhance prompts for AI-generated code, products, or documents. Triggers include requests to "optimize my prompt", "improve this requirement", "make this more specific", or when raw requirements lack detail and structure.
Local-first codebase intelligence for AI assistants via MCP
Repository knowledge engine plugin. Bundles CLI-backed slash commands (rb-setup / rb-init / rb-refresh / rb-ask) plus the agent-repo-init skill. Works in both Claude Code and Codex CLI.
Repository knowledge engine plugin. Bundles CLI-backed slash commands (rb-setup / rb-init / rb-refresh / rb-ask) plus the agent-repo-init skill. Works in both Claude Code and Codex CLI.
信念对决。当父子节点内容冲突、或两条你都认可的记忆逻辑上不能并存时使用。
记忆审计入口。当我主动决定审视记忆质量时，先读此文件判断应使用哪个子技能。
Advanced PM skills: AI Product Canvas, Multi-Source Signal Synthesiser, Experiment Designer, Design Handoff Brief. For senior PMs working on complex or AI-powered products.
可发现性审计。当disclosure写法有问题、parent放错、alias缺失、子节点过多时使用。
模式提取与失效解药分析。当发现多条记忆在讲同一个教训，或发现自己在一而再再而三地犯同样的错误时使用。
Data & metrics skills: Data Analysis Standard, Retention Analysis, Product Health Analysis. Structure metric deep-dives, funnel analysis, cohort studies and churn investigations.
死数据清洗。当一条记忆读不读你的行为都不会变、感悟没有现实锚点时使用。
Claude Code integration for LycheeMem structured long-term memory.
Forceful operating rules for using LycheeMem as the primary structured long-term memory path inside OpenClaw.
A tool to vectorise repositories for RAG.
Methodology for writing or improving prompts and system prompts that drive any LLM. Use when authoring or revising a prompt for a model task — grouping, classification, extraction, generation, copywriting, labeling, agent instructions, prompt templates, skill instructions — to decide how much to constrain the model based on the task type (open-ended vs single-correct-answer) and write the most fitting instructions. Triggers: "write a prompt", "help me write or improve a prompt", "how should I change this prompt", "this prompt isn't working", "write instructions for the model", "prompt-writer". Not for: answering the user's question directly, or writing articles and documents meant for human readers (those are not prompts that drive a model).
Biologically-inspired memory for Claude Code. Memories decay by default, retrieval strengthens them, errors stick longer. Auto-captures context at session start and learns from errors.
Self-improving Claude Code and Codex plugin that turns corrections into Preferences, Project-specific skills, and Shared skills
Use SwarmVault when the user needs a local-first knowledge vault that writes durable markdown, graph, search, dashboard, review, chat-session, context-pack, task-ledger, static AI export, retrieval, and MCP artifacts to disk from books, notes, transcripts, exports, datasets, slide decks, files, URLs, code, and recurring source workflows.
Add your description here
Add your description here
Unified inner work engine: Schema deconstruction (diagnosis) + IFS therapy (treatment). Absorbs: schema-deconstruction.
Activated for marketing campaigns, content strategy, distribution planning, SEO, and audience growth. Deploys the marketing swarm when full execution is needed.
Opinionated agentic RAG powered by LanceDB, Pydantic AI, and Docling
Unified zero-emotion variance shield for capital markets (FX, crypto, CFDs) and poker. Absorbs trading-risk-gate + zenith-execution + trade-journal-analyzer into one engine.
Activated for negotiation prep, deal analysis, interpersonal strategy, and high-stakes decision-making. Combines game theory with psychological awareness.
Use when reviewing a proposed REST or GraphQL API change before merge — checks contract clarity, backwards compatibility, errors, pagination, auth, and naming.
Unified boundary enforcement, interpersonal diagnostic, and relational audit engine. Absorbs 40 psychology + 2 social protocols and all relationship case studies.
Activated for trading analysis, position sizing, risk management, and trade structure discussions. Preloads all trading-domain skills for complete analytical capability.
Use when first encountering a new dataset — produces a structured profile (schema, missingness, distributions, outliers, gotchas) before any analysis.
PROACTIVELY activated when any discussion involves risk, ruin, leverage, large commitments, or irreversible decisions. Auto-invokes safety gates before analysis proceeds.
Use after an incident is resolved — drafts a blameless postmortem from timeline notes, alerts, and chat threads.
Use when capturing an architecture decision so it survives turnover — produces an ADR-NNNN.md from context, options considered, and the chosen path.
Claude Mind - Give Claude photographic memory in ONE portable file. Share, version, and transfer your Claude's brain.
Use after a session to promote useful episodic notes from logs/episodic/ into distilled, dated entries in MEMORY.md and memory/semantic/.
Use when the user asks for a sourced briefing on a topic that spans multiple web sources and requires citations.
Persistent semantic memory for Claude Code — auto-injects relevant context at session start, captures git commits, test results, and research via hooks, and provides 30+ MCP tools for memory management.
Use before opening a PR to audit the changes for stale comments, unused imports, missing tests, and inconsistencies with neighboring code.
Use when planning the next sprint — turns ticket intake + team capacity into a planned sprint with explicit non-goals.
Semantic search for Claude Code conversations. Remember past discussions, decisions, and patterns.
Use when opening a PR — produces a clean PR description (what / why / how to verify / risks) from a branch diff against base.
Use before connecting a new MCP server to your agent — produces a structured security review covering source, permissions, tools, network, and approvals.
Python SDK for Ratel — context engineering platform for AI agents. BM25 tool retrieval, MCP ingestion, framework-neutral capability tools.
AI-powered code indexer with semantic search and knowledge graphs
Use when the user wants a quality review, interaction audit, or to test the workflow against realistic scenarios.
Use when the workflow lacks error handling, has been failing in production, or needs retry logic, fallback strategies, and circuit breakers.