Preview. These maps are our reading of public sources and haven't been checked with the projects' maintainers yet.
Skills and plugin packs for coding agents
Collections of agent skills, plugins, commands, and prompts that people publish for coding agents.
20 projects75% active100% build on a shared standardMapped 2026-10-26
What they agree on
- Is it actively maintained? active (75% of 20)
- How is the main thing licensed? permissive open source (MIT/Apache/BSD) (85% of 20)
- Can it run entirely on your own device or infrastructure, without the maker's servers? yes (95% of 20)
- Do you need an account to use it? no account (95% of 20)
- Who is it for? developers (80% of 20)
- Which shared standard or format does it mainly build on? Agent Skills (SKILL.md) (75% of 20)
- Does it need an outside service beyond the agent and its model? no (75% of 20)
- Does it send usage data to its maker? none documented (85% of 20)
What they argue about
How is it mainly sustained?
Which coding agents does it give install steps for?
What is most of it, whatever it's packaged as?
How is it mainly installed?
How big is it?
What work is it for?
How does it get the agent to follow it?
Where does its know-how mainly come from?
Does it publish evidence that it makes agents better?
Projects
| Project | Status | What it is |
|---|---|---|
| agent-skills (implement-spec) | Active | Harness-agnostic Agent Skills for rigorous engineering: implement-spec runs a spec to a verified PR, plus review, review-feedback, and docs… |
| AI-Native Boilerplate | Maintenance | A project boilerplate with 160+ layered rules, 16 lifecycle hooks, and git hooks to keep AI coding agents on spec, for Claude Code, Cursor,… |
| Anthropic Agent Skills (anthropics/skills) | Active | Anthropic's public repository of example and document skills (docx, pdf, pptx, xlsx and more) for Claude, installable as a Claude Code plug… |
| anti-ai-slop | Active | One writing skill that rewrites a draft to remove AI-writing tells, then runs it through four gates: invisible characters, a Python linter,… |
| book-skills | Maintenance | Claude Code skills, one per book (Lean Startup, Pragmatic Programmer, K&R C and others), that apply each book's framework when its ideas co… |
| Chess post-mortem skills | Active | Four Claude Code skills that analyze one of your chess games with Stockfish and your own think-aloud notes, producing an annotated PGN, HTM… |
| Claude Skills Starter Kit | Maintenance | 20 free domain skills (agents, RAG, GEO, security, product) for Claude Code and Antigravity, as a teaser for a paid 198-skill pack. |
| Deslop | Active | A single portable skill that rewrites legalese and AI slop into plain English using Bryan Garner's plain-language rules, for lawyers using… |
| don't let me (mentor plugin) | Maintenance | A plugin that turns your coding agent into a mentor that knows your goals and checks your work against a 'don't let me' list. |
| frontend-design-pro | Active | A frontend UI/UX skill pack with a small router skill that lazy-loads 19 catalog skills, plus TypeScript checkers and 11 release gates. |
| HumanLayer show-me | Active | A single skill from HumanLayer that makes a coding agent explain things with compact diagrams, code sketches, and small HTML artifacts. |
| Kage | Active | A design inspiration library (587 designs, 3,096 components) that turns real product UIs into prompts, served to agents through a hosted MC… |
| LinkedIn skills | Active | 12 Claude Code and Codex skills that draft LinkedIn posts and comments in your voice, analyze your feed, and schedule posts after your appr… |
| OpenCode Senses | Active | An OpenCode plugin that gives text-only coding models vision (OCR, object location, colors, image search) with a local Moondream model. |
| Poka-Yoke | Active | Shigeo Shingo's mistake-proofing method as 11 skills, a hazard scanner, and installable pre-commit/CI/hook devices, benchmarked across mode… |
| SuperCMO Skills | Active | Open-source skills and an MCP server that let agents produce marketing campaigns (UGC videos, ad videos, product photos) from a product pho… |
| Superpowers | Active | Jesse Vincent's skills framework and development methodology (brainstorm, plan, subagent-driven development, TDD) installed as a plugin in… |
| Trail-code (formerly Graft) | Changed direction | A codebase map plus team 'learnings' fed to Claude Code, Cursor, Codex and other agents through hooks, to cut search tokens and tool calls. |
| Whetstone | Active | A set of 31 short Claude Code skills, each distilled from one real failure, grouped into verify, plan, change safely, diagnose, and author. |
| wshobson/agents | Active | Seth Hobson's multi-harness plugin marketplace: 94 plugins with about 202 subagents, 185 skills, and 105 commands for Claude Code, Codex, C… |
Where they meet
- awesome-claude-code awesome list
~57k stars, pushed 2026-10-26; LinkedIn skills shows its 'Mentioned in' badge. - Claude Code official plugin directory marketplace (GitHub)
Anthropic-managed; ~45k stars; Superpowers installs from it. - OpenAI Codex plugins marketplace (GitHub)
Official Codex plugin marketplace; lists Superpowers. - Agent Skills specification standard and docs
Spec repo github.com/agentskills/agentskills (~27k stars). - skills.sh directory and installer (npx skills)
'The Agent Skills Directory'; used by deslop, frontend-design-pro, HumanLayer, wshobson. CLI is vercel-labs/skills. - awesome-agent-skills awesome list
1000+ skills across agents; ~36k stars, pushed 2026-10-25.
Follow
RSS for this community: projects added, status changes, and later polls and discussions.
Follow the whole community: OPML of every project's own release feed (where we found one) · OPML of every project's Poll Position feed.
The poll
Every project above was answered against these questions. Core questions are shared by all communities; the rest are this community's own.
- Is it actively maintained? core
active · active, but changed direction (pivot, fork, or rename) · maintenance or slowing · dormant, sunsetting, or archived - How is the main thing licensed? core
permissive open source (MIT/Apache/BSD) · copyleft or source-available with restrictions · open parts around a closed core · proprietary or closed source - How is it mainly sustained? core
individual side project, no revenue · individual with revenue (paid app, tier, ads, or donations) · community, foundation, or research · company with a paid product, tier, or ads · company or startup with a free product (no revenue yet) · big company side project - Can it run entirely on your own device or infrastructure, without the maker's servers? core
yes · yes, but it calls a third-party service you choose (your own API key or account) · partly (some features need the maker's servers) · no, it needs the maker's servers · not applicable - Do you need an account to use it? core
no account · optional account · account required - Who is it for? core
developers · non-developers · both - Is there public evidence it's built with AI coding agents? core
yes, the repo shows it (agent instruction files or agent co-authored commits) · yes, the maker says so · no public evidence - Which shared standard or format does it mainly build on? core
Agent Skills (SKILL.md) · an agent's plugin or hooks format (Claude Code plugins and hooks, OpenCode plugins) · Model Context Protocol (MCP) · AGENTS.md or agent rules files (CLAUDE.md, .cursorrules) · none of these (its own format or interface) · not applicable - Which coding agents does it give install steps for? approach
one agent · 2 to 5 named agents · 6 or more named agents, or any agent via a cross-agent installer (npx skills) - What is most of it, whatever it's packaged as? approach
Agent Skills (SKILL.md folders), even if packaged as a plugin · a mix where subagents, commands, or hooks matter as much as skills · code that runs alongside the agent (a plugin, hooks, or an MCP server) · rules or instruction files (CLAUDE.md, AGENTS.md, .cursorrules) - How is it mainly installed? approach
a plugin marketplace command · a package manager or installer CLI (npm, npx skills, pip) · clone or copy the files by hand · connect to a hosted service - How big is it? approach
one skill or tool with one job · a small set (2 to 10 skills or commands) · a large bundle (more than 10) - What work is it for? approach
general coding workflow and discipline · one technical area (frontend, security, codebase search, vision) · work outside coding (writing, marketing, law, chess, coaching) · a mix across many areas - Does it need an outside service beyond the agent and its model? approach
no · optional (some features use outside services or keys) · yes, it requires an outside service, account, or key - How does it get the agent to follow it? value
instructions only: the model decides (a hook that only loads instructions counts here) · instructions plus scripts or checks the agent is told to run · hooks or gates that block or check work automatically, outside the model's choice · not applicable (it adds tools or context, not rules to follow) - Where does its know-how mainly come from? value
the maker's own practice and failures · published works, guidelines, or methods · data mined at scale (posts, designs, benchmarks) · a vendor documenting its own product · not applicable (it adds tools or indexes your code, not know-how) - Does it publish evidence that it makes agents better? value
yes, evals or benchmarks with numbers · an eval harness or tests, but no published results · examples or demos only · no - Does it send usage data to its maker? value
none documented · on by default, with an opt-out
Sources
Show HN posts 2026-07-27 to 2026-10-25 in this kind, plus well-known reference projects.
27 sources across 20 projects, read on 2026-10-26. Each project page lists its own. Poll version: core-v4.