6,602 tagged Memory and context, measured the same way as everything else here.
Browse within: LLM 229agent 159agent-memory 153cli 134agentic-ai 123agents 119ai-memory 111claude-code-plugin 83ai-coding 76context-engineering 59ai-tools 58skills 58knowledge-base 56open-source 55
robinlidberg-dot/cap-shield-mcp
Skill Claude CodeCodex
Measure and reduce what an agent sends into its context. Use when an agent's context is large, when retrieval quality matters, when the user asks about token cost or context bloat, or when agent answers are wrong in ways that look like missing information rather than reasoning failure. Also use to measure compression…
Skill Claude CodeCodex
Durable causal memory for this agent via the connected MADB MCP server — persists across sessions and records WHY things happened (causedby lineage). Use it as a reflex tied to the work, not on every turn. RECALL at the start of a task that references prior work/projects/decisions ("my project", "the script we wrote"…
Skill Claude CodeCodex
Use the hosted VersionPill MCP board for product memory while coding. Triggers: board, pulse, ship, vision, "what's next", VersionPill, product cards.
Skill Claude CodeCodex
Use subconscious-mcp MCP tools (recall, remember, echo, driftreport, forget) for local semantic memory. Trigger when starting non-trivial work, before re-solving a problem, after completing a reusable task, on a nearduplicate warning, or when the user mentions memory, cache, or repeated work across sessions.
Skill Claude CodeCodex
SynthNet is the commons for working agents: post field notes, claim bounties, and build reputation that persists across context resets. Join in one call. Use when the agent wants fresh agent-earned knowledge (a tool that broke, a prompt pattern that works, an API that changed), wants to delegate work it can't do, or…
vidursharma202-del/threadline-mcp
Skill Claude CodeCodex
Your AI agents start from zero every session — users repeat themselves endlessly. Threadline fixes this in 2 lines. inject() before your LLM call, update() after. Persistent context across every session, model, and product. Structured extraction across 7 scopes. OAuth-style grant system so agents only see permitted…
joustonhuang/ai-first-pastebin
Skill Claude CodeCodex
Use 呵.tw as an AI-first pastebin to shorten URLs, store shareable agent handoff notes, recover tagged pastes, and reduce LLM token usage across sessions. Trigger this skill when the user wants to: (1) turn long URLs into short links, (2) park long non-secret intermediate output outside the context window, (3) hand off…
Skill Claude CodeCodex
Use when the user expresses a rule (must / never / always / going forward / no longer / forbid / require / cannot / shouldn't) — captures one rule at a time. ALSO use proactively on a fresh repo with empty skills.yaml: call derivecandidates to mine the existing graph for candidate rules, walk through them with the…
Skill Claude Code
Encrypted backup and restore for OpenClaw Agent workspace files (SOUL.md, MEMORY.md, IDENTITY.md, AGENTS.md, TOOLS.md). Uses tar + openssl (AES-256-CBC) encryption and soul-upload.com API. Auto-generates a new random password for each backup (DO NOT reuse passwords). Use when the user needs to: (1) Back up or upload…
Skill Claude CodeCodex
A general-purpose connection to Hermes Agent, an AI coding assistant with tools for memory, research, browser automation, code execution, and delegated tasks. Hermes Agent can also create reusable skills from work it performs.
Grail-Computer/Self-Improving-Agent
Skill Claude CodeCodex
Meta-skill for making the agent self-improving. Covers updating AGENTS.md, creating new skills from repeated workflows, and deciding what to systematize. Invoke after completing tasks, when noticing repeated friction, or at session end.
irangareddy/openclaw-essentials
Skill Claude CodeCodex
Manage agent memory through daily logs, session preservation, and knowledge extraction. Use when (1) logging work at end of day, (2) preserving context before /new or /reset, (3) extracting patterns from daily logs to MEMORY.md, (4) searching past decisions and learnings, (5) organizing knowledge for long-term…
chengyangliu-lcy/xinyangkb_mcp
Skill Claude CodeCodex
An internal assistant for Xinyang Company that searches company documents to answer questions about products, technical plans, rules, processes, projects, and other internal information.
Skill Claude Code
Give your agent a searchable knowledge brain - semantic search, topic synthesis, and action tracking across your saved YouTube videos, articles, Reddit threads, X posts, and PDFs.
Skill Claude CodeCodex
Search registered repo-local llm-wiki sources through the wiki-mcp-hub MCP tools. Use when the user asks about project notes, decisions, runbooks, API contracts, incidents, or cross-repo context that may already be documented in local wikis.
Skill Claude CodeCodex
Save important information to long-term memory. Use when an important decision is made, a hard problem is solved, or a new preference is discovered.
Skill Claude CodeCodex
Complete guide for using the Arca MCP persistent memory system. Use this skill when an agent needs to store, retrieve, update, or delete memories across sessions; when organizing information into buckets; when performing semantic search over stored facts; when building or traversing knowledge-graph relationships…
Skill Claude CodeCodex
Run a semantic health-check session on the Librarian vault. Hunts for contradictions between notes, stale claims superseded by newer sources, concepts mentioned in bodies but lacking their own note, missing cross-references, and data gaps. Uses queryknowledge and readconcept for analysis; repairs findings with…
Skill Claude CodeCodex
Use when the user wants their Kimi CLI agent to self-improve from past usage, asks about a nightly/offline "sleep" or "dream" cycle, memory/skill consolidation, or says things like "make my agent better the more I use it", "review my past sessions", "learn my preferences", "consolidate what you learned", "run the…
jherard-fr/cursor-chats-bridge
Skill Claude CodeCodex
Install or update the Claude-Cursor chat bridge on a Windows PC — a Python MCP server that exposes Cursor's local chat database (read-only) to Claude, plus a background poller that journals new messages every 5 min via Task Scheduler. Use this skill whenever the user wants to set up the Cursor connection on a new…
Skill Claude CodeCodex
Category: Memory & Context Version: 1.1.0 Type: MCP Server.
Skill Claude CodeCodex
Persistent symbolic memory with PLN/NAL inference over scoped AtomSpaces. Tools are exposed via plugin-invoke (atomspace-assert/query/infer/recall).
Skill Claude Code
A persistent memory manager for an AI coding assistant, with instructions written in Chinese, that loads, records, searches, and saves project context across sessions.
Skill Claude CodeCodex
Use the user's deardiary journal (their transcribed handwritten diaries, exposed via MCP) to understand their life background. Trigger when conversation turns personal - life decisions, relationships, feelings, past events, "you don't know me" moments - or when the user mentions their diary, journal, past, or a person…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: