tech-leads-club/agent-skills is a TypeScript skill registry that hit GitHub trending on 2026-05-17 with +923 stars in 24 hours (3,384 total) — making it one of the fastest-growing entries in the May 2026 skills-as-runtime convergence wave. Unlike vendor-specific bundles, it ships skills as strictly-typed TypeScript modules that load into Claude Code, Cursor, Antigravity IDE, and GitHub Copilot through a thin compatibility shim per harness. The bet is that as the SKILL.md format matures, the bundle authors that win will be the ones offering schema-checked, type-safe skills rather than free-form markdown. Pairs with addyosmani/agent-skills (mainstream DX), mattpocock/skills (Claude Code-first), and ComposioHQ/awesome-codex-skills (Codex-first) to cover the four-harness landscape with a typed-registry angle.
tech-leads-club/agent-skills is a TypeScript-typed registry of agent skills targeting four harnesses simultaneously: Claude Code, Cursor, Anthropic’s Antigravity IDE, and GitHub Copilot. The repo crossed 3,384 stars on 2026-05-17 (+923 in 24h), placing it among the fastest-rising entries in the skills-as-runtime category for May 2026.
The differentiator is types. Most skill registries publish free-form SKILL.md files; this one wraps each skill in a typed TypeScript module that defines explicit input/output contracts, capability surface, and harness compatibility. A schema-validation layer rejects malformed skills at install time, which makes the registry a candidate for teams that want to ship internal skills across multiple harnesses with the same confidence guarantees they’d apply to any other package dependency.
The skills-as-runtime trend has been dominated by markdown-first bundles. tech-leads-club/agent-skills is a credible vote for the typed-registry approach — useful for teams that already invest in TypeScript-first internal tooling and want skill governance to mirror that. The fact that it hit GitHub trending on launch week alongside addyosmani’s bundle is a signal that the market is testing both formats in parallel rather than converging on one.
AI agents that work well with tech-leads-club/agent-skills.
Production-grade engineering skills for AI coding agents — addyosmani's curated bundle covering Claude Code, Cursor, and Antigravity workflows.
100+ ready-to-use Claude Code agents, commands, hooks, and integrations — installable via CLI or interactive dashboard at aitmpl.com.
Injects up-to-date, version-specific documentation into LLM context — eliminates hallucination from stale training data.