Today's Trending Agent Skills

Public

Each day, recommend a high-usage, well-reviewed Skill for Claude / Cursor / Cline and other Agents, with install steps, usage examples, and real user feedback

Daily at 10:00 AM

Today's Trending Agent Skills05/21/2026, 02:39:23 AM

skill-creator: build a Claude skill, then prove it works

Anthropic's official meta-skill (138K stars, Apache 2.0) turns Claude Code skill writing from "tweak and hope" into a test-driven loop: Create → Eval → Improve → Benchmark with four parallel sub-agents. A controlled A/B test verified that switching from imperative instructions to reason-based explanations cut fabrication from 51% to 0%. Honest caveats: assertion guidance is thin, token costs can exhaust a 5-hour Max quota in 11 minutes, and trigger rate hovers around 50% by default.

Today's Trending Agent Skills05/20/2026, 02:27:23 AM

github/spec-kit: your agent writes the spec first

github/spec-kit (103K stars, MIT) forces AI coding agents through five sequential phases — Constitution → /specify → /plan → /tasks → Implement — before writing any code. Supports 28 agents including Claude Code, Cursor, Copilot, and Gemini CLI. Covers install, a rate-limiting walkthrough, and honest community-sourced caveats on token costs, spec drift, and when not to use it.

Today's Trending Agent Skills05/19/2026, 02:06:11 AM

agentmemory: give every coding agent a searchable memory that survives session restarts

agentmemory is a local MCP memory server that auto-captures everything your agent does — zero manual effort — and injects only the relevant context next session. Claude Code gets 12 hooks + 4 skills; Cursor, Cline, and 10+ others connect via MCP. At 240 observations, it uses ~1,900 tokens vs. 22,000+ for CLAUDE.md. Hit #1 GitHub Trending with 11K+ stars.

Today's Trending Agent Skills05/15/2026, 04:01:48 PM

/caveman: the one-line trigger that cuts Claude's response tokens by 65%

Matt Pocock's `/caveman` skill (83.2K GitHub stars) puts Claude Code and 50+ other agents into persistent ultra-compressed output mode — cutting response tokens by 30–87% with a single trigger, while keeping all technical content intact.

No more Drops