new file mode 100644
index 0000000..fc4bee7
@@ -0,0 +1,58 @@
+---
+visibility: public
+---
+
+# tooling & infrastructure
+
+skill creators, installers, marketplaces, and orchestrators — the meta-layer.
+
+## creating skills
+
+- **[skill-creator](https://github.com/anthropics/skills/tree/main/skills/skill-creator)** — Anthropic's official meta-skill. includes eval framework and description optimizer. see [[creating-skills]].
+- **[Skill Seekers](https://github.com/yusufkaraaslan/Skill_Seekers)** — auto-converts any documentation website into a Claude skill. by @yusufkaraaslan
+- **[claude-code-skill-factory](https://github.com/alirezarezvani/claude-code-skill-factory)** (685 stars) — toolkit for building production-ready skills.
+
+## installers & package managers
+
+- **[antigravity](https://github.com/sickn33/antigravity-awesome-skills)** (32.3k stars) — 1,370+ installable skills with CLI installer. works across Claude Code, Codex, Gemini CLI, Cursor.
+- **[CCPI](https://github.com/jeremylongshore/claude-code-plugins-plus-skills)** (1.9k stars) — 340 plugins + 1,367 skills with package manager.
+- **[skillshare](https://github.com/runkids/skillshare)** (1.4k stars) — sync skills across all AI CLI tools with one command.
+- **[open-skills](https://github.com/instavm/open-skills)** (398 stars) — run Claude Skills locally using any LLM.
+
+## marketplaces & directories
+
+| marketplace | notes |
+|-------------|-------|
+| [SkillsMP](https://skillsmp.com) | aggregates from GitHub |
+| [Claude Marketplaces](https://claudemarketplaces.com) | 2,400+ skills, 2,500+ marketplaces |
+| [skills.pawgrammer.com](https://skills.pawgrammer.com) | 280+ free community skills |
+| [buildwithclaude.com](https://buildwithclaude.com) | plugin marketplace |
+| [claudepluginhub.com](https://www.claudepluginhub.com) | plugin directory |
+| [AITMPL](https://www.aitmpl.com/plugins/) | 497+ extensions |
+| [OfficialSkills.sh](https://officialskills.sh) | 300k+ monthly views |
+
+## orchestrators & multi-agent
+
+- **[claude-squad](https://github.com/smtg-ai/claude-squad)** — manage multiple Claude agents in separate workspaces.
+- **[claude-swarm](https://github.com/parruda/claude-swarm)** — agent swarms.
+- **[claude-code-flow](https://github.com/ruvnet/claude-code-flow)** — code-first orchestration.
+- **[claude-octopus](https://github.com/nyldn/claude-octopus)** (2.5k stars) — puts up to 8 AI models on every coding task.
+- **[claude_code_agent_farm](https://github.com/Dicklesworthstone/claude_code_agent_farm)** (780 stars) — orchestration for 20+ agents in parallel.
+
+## memory & context
+
+- **[claude-mem](https://github.com/thedotmack/claude-mem)** (48.2k stars) — captures everything Claude does, injects context into future sessions.
+- **[beads](https://github.com/gastownhall/beads)** (20.6k stars) — memory upgrade for coding agents.
+- **[claude-supermemory](https://github.com/supermemoryai/claude-supermemory)** (16.7k stars) — persistent memory, tops benchmarks (LongMemEval, LoCoMo, ConvoMem).
+- **[Acontext](https://github.com/memodb-io/Acontext)** (3.3k stars) — Agent Skills as a Memory Layer.
+- **[second-brain-skills](https://github.com/coleam00/second-brain-skills)** (664 stars) — turns Claude Code into a Second Brain.
+
+## observability
+
+- **[claude-hud](https://github.com/jarrodwatts/claude-hud)** (18.4k stars) — shows context usage, active tools, running agents, todo progress.
+- **[multi-agent-observability](https://github.com/disler/claude-code-hooks-multi-agent-observability)** (1.4k stars) — real-time monitoring via hook event tracking.
+
+## writing output
+
+- **[caveman](https://github.com/JuliusBrussee/caveman)** (18.6k stars) — cuts 65% of tokens by making Claude talk like a caveman. genuinely useful for reducing cost/context consumption.
+- **[humanizer](https://github.com/blader/humanizer)** (13.4k stars) — removes signs of AI-generated writing.
\ No newline at end of file