Malpiedi

Agent Radar Yearbook — Development skill for Claude Code

Development community

Rank the top repos of a whole year in the Claude skills, subagents and MCP ecosystem - scored on momentum, trust and relevance, then sorted into subject domains.

How to install Agent Radar Yearbook

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Malpiedi/agent-radar-yearbook and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Agent Radar Yearbook does

Rank the top repos of a whole year in the Claude skills, subagents and MCP ecosystem - scored on momentum, trust and relevance, then sorted into subject domains. Year-in-review companion to agent-radar.

Alternatives in Development

Internet Court Skill — The trust layer for agent-to-agent commerce — natural-language mandates, ERC-7710 delegated permissions, x402 4.9k ★
Om Meeting — Prep for any meeting by topic — subject-forward briefing with open tasks, blockers, and brainstormed considera 4.6k ★
CORS — Scan an endpoint for CORS misconfiguration — arbitrary-origin reflection, null-origin trust, credential exposu 4.5k ★

README

Agent Radar Yearbook

A Claude Code skill that ranks **the top repos of a whole year** in the Claude skills, subagents, plugins and MCP ecosystem — and files each one under what it is actually for.

Companion to [agent-radar](https://github.com/Malpiedi/agent-radar), which does the same job over a rolling window of days or weeks. Same three scoring axes, same thresholds, one deliberate difference in how momentum is measured.

node year.mjs

Output: a ranked table of the top 50, a domain summary, the entries grouped by domain, then the repos it kept out of the ranking and why.

# Agent radar yearbook — 2026

Repos created 2026-01-01 → 2026-09-10 · min 10 stars · categories: skills, mcp, list
20 queries × 3 windows → 1,867 unique candidates, 250 enriched, 744 API calls.

| # | Repo                        | Domain                | Stars |   /day |  M |  T |  R | Total |
|---|-----------------------------|-----------------------|-------|--------|----|----|----|-------|
| 1 | vercel-labs/agent-browser   | Testing & QA          | 42.3k | 174.84 | 40 | 38 | 20 |  98   |
| 2 | alibaba/open-code-review    | Coding & code quality | 22.2k | 192.47 | 38 | 40 | 20 |  98   |
| 3 | mvanhorn/last30days-skill   | Research & web        | 61.7k | 268.96 | 39 | 38 | 20 |  97   |

What it checks

	out of	signals
Momentum	40	stars/day across the repo's life, days since last push, days between creation and last push
Trust	40	owner account age, followers, contributor count, license, README size, CI, tests, org ownership, fork ratio
Relevance	20	what is in the git tree: `SKILL.md` count, `.claude/agents/`, plugin marketplace, MCP manifest

Relevance gates entry on purpose. A repo with 30,000 stars that merely mentions agents in its README cannot enter the ranking over a smaller one that actually ships skills.

Red flags are tracked separately rather than subtracted, and blocking ones keep a repo out of the