Best memory & context skills for agents
A ranked comparison of memory & context skills, MCP servers and repos for AI agents: 15 of 611 passed our security scan. Ranked by curation, then GitHub stars. No paid placement.
- 1. Superpowers (obra) Skill, 301k GitHub stars
- 2. Matt Pocock Skills (handoff, research, grill-me) Skill, 295k GitHub stars
- 3. Mem0 Repo, 67k GitHub stars
ItemScanGitHub starsLicense
Superpowers (obra)Repo pass301kMIT Matt Pocock Skills (handoff, research, grill-me)Repo pass295kMIT Mem0Pass67kApache-2.0 Obsidian skillsRepo pass50kMIT Task Master (claude-task-master)Repo pass28kNOASSERTION Beads (issue memory for agents)Pass28kMIT NangoPass13kNOASSERTION Notion MCP server (official)Repo pass4.7kMIT Task chunking rule (no job over 40 minutes) (Avatar 8)Pass–Avatar 8 Venture Operator guardrails (Avatar 8)Pass–Avatar 8 mem0-vercel-ai-sdkRepo pass67kApache-2.0 composioRepo pass31kMIT stacklok/toolhivePass2.3kApache-2.0 toolhiveRepo pass2.3kApache-2.0 parallel-memoryRepo pass84MITPopular but not yet scanned (12)
- morluto/rea 307k GitHub stars, not yet security-scanned
- reverse-engineer-anything 307k GitHub stars, not yet security-scanned
- affaan-m/ECC 282k GitHub stars, not yet security-scanned
- ECC 282k GitHub stars, not yet security-scanned
- memory-leak-audit 194k GitHub stars, not yet security-scanned
- JavaGuide 159k GitHub stars, not yet security-scanned
- Snailclimb/JavaGuide 159k GitHub stars, not yet security-scanned
- component-refactoring 159k GitHub stars, not yet security-scanned
- langgenius/dify 159k GitHub stars, not yet security-scanned
- academic-researcher 142k GitHub stars, not yet security-scanned
- Shubhamsaboo/awesome-llm-apps 142k GitHub stars, not yet security-scanned
- graphify 128k GitHub stars, not yet security-scanned
Section: Memory & context. For agents: full markdown table.