Migrating from awesome-claude-code to ClaudSkills
If you've been finding Claude Code skills via awesome-list READMEs on GitHub — awesome-claude-code, awesome-anthropic, or any of the curated lists that appear when you search "Claude skill" — here's a practical guide to using ClaudSkills alongside or instead of them. Not a takedown of awesome-lists; they're great. Just a different shape that solves different problems.
What awesome-lists do well
The awesome-list format earned its popularity for good reasons. It's lightweight (just a README), trustworthy (the maintainer's personal pick), browsable on phones via GitHub mobile, and most importantly — every entry is human-vouched by a real maintainer who knows the ecosystem. If you trust the curator's taste, you can find a great starting set in ~5 minutes.
What they don't do well, by design:
- Updates lag what's new — most are weekly or monthly cadence.
- No search beyond Ctrl+F in a markdown file.
- No category or tag filtering — you scroll through every section.
- No programmatic access — the data is in a markdown file's prose, not a queryable schema.
- Coverage is bounded by what the maintainer has time to vet — usually <500 entries.
What ClaudSkills does well
ClaudSkills started as a complement, not a replacement. The architecture optimizes for different things:
- Volume and freshness. 69,016 admitted skills today (vs ~100-500 in any single awesome-list), mined twice daily from 26 public sources including GitHub Topics, code search, awesome-list READMEs themselves, Reddit, HN, Bluesky, Mastodon, gists, dev.to, marketplaces, newsletters.
- Faceted search. Filter by 10 categories, 118 sub-categories, 58 orthogonal tags (language, framework, cloud, AI provider, type). Each filter is a real URL — bookmarkable, shareable, indexable.
- Programmatic dataset. Full catalog at
/data/skills.json— ~5 MB JSON, refreshed nightly. Build whatever indexer, recommender, or pipeline you want. - Content-derived ranking. Pro subscribers see a multi-signal Quality Score (80% structural rubric + 20% metadata depth). Free public payload carries a
daily_eligibleboolean for the top-12%. - Per-skill OG cards + permalink stability. Every /skills/<slug>/ page has a unique social-share card; the URL never moves.
- No paid placement, no listing fees. Admission is rule-based on content quality alone.
Practical migration paths
Path 1: "I was just browsing for inspiration"
Replace your awesome-list bookmark with one of these ClaudSkills entry points:
- /awesome/ — Top 100 standout skills (mirrors the curated-list feel)
- /best/ — Hand-curated for production use
- /sotd/ — One curated pick per UTC day, with an RSS feed and the same pick on the mobile companion app
- /recent/ — Newest mined skills (last 7 days) — with its own RSS feed
Add the RSS feeds to Feedly / dev.to / Inoreader. New skills land in your reader without manual checking.
Path 2: "I was looking for a specific use-case"
ClaudSkills has 10 hand-curated use-case bundles that group skills by job-to-be-done rather than by category. Map your awesome-list section headers to these URLs:
| If you were looking under… | Go to |
|---|---|
| "Security" / "Audits" | /use-case/security-audit/ |
| "Code review" | /use-case/code-review/ |
| "Content / Writing" | /use-case/content-writing/ |
| "Cold email / Outreach" | /use-case/cold-outreach/ |
| "Data analysis" | /use-case/data-analysis/ |
| "DevOps / Infrastructure" | /use-case/devops-automation/ |
| "Customer support" | /use-case/customer-support/ |
| "SEO / Marketing" | /use-case/seo-research/ |
| "Social media" | /use-case/social-media/ |
| "Testing" | /use-case/test-generation/ |
Path 3: "I was building an indexer / aggregator that scrapes catalogs"
Don't scrape ClaudSkills — fetch /data/skills.json directly. The full catalog (slug, name, description, category, sub-category, tags, daily_eligible flag, author, license, source URL) is a single JSON file.
# Python — fetch + filter to Sales/security daily_eligible skills
import json, urllib.request
raw = urllib.request.urlopen('https://claudskills.com/data/skills.json').read()
catalog = json.loads(raw)['skills']
elite = [s for s in catalog
if s.get('daily_eligible')
and s.get('category') in {'sales', 'security'}]
print(f'{len(elite)} daily-eligible skills in sales+security')For per-skill raw markdown, fetch /skills/<slug>/SKILL.md — served via Cloudflare R2 with appropriate caching headers.
Path 4: "I was an awesome-list maintainer myself"
Two paths that don't conflict:
- Keep maintaining your list. ClaudSkills' miner crawls awesome-list READMEs as one of its 26 sources, so anything you curate gets ingested automatically. Your hand-vouched picks bubble up to higher quality_pro scores anyway because awesome-list inclusion correlates with structural authoring discipline.
- Submit your list's standouts directly. The Submit form takes a GitHub URL pointing at a SKILL.md file. Useful for new skills you want indexed within ~12 hours rather than waiting for the next awesome-list-README crawl pass.
Honest caveats
ClaudSkills isn't a strict superset of awesome-lists. Some real trade-offs:
- Curation taste. ClaudSkills admission is rule-based — anything that clears the structural-quality threshold is in. A good awesome-list maintainer applies subjective taste that no rubric can match. If your decision criterion is "what would Sindre Sorhus recommend for production use", an awesome-list is the better fit.
- Editorial collections. The few editorial selections we maintain (/best/, /awesome/, /sotd/) are smaller and less opinionated than a full awesome-list.
- You may already have your patterns down. If you're a power user who has bookmark-managed your way through awesome-lists for years, you have working muscle memory. ClaudSkills has different muscle memory — it pays off most when you're searching faceted-ly or when you need programmatic access.
How awesome-list maintainers can crosswalk content
If you maintain an awesome-list and want to add a "browse the full catalog" footer pointing at ClaudSkills, this is the canonical URL pattern:
## Browse the full catalog
For programmatic access or filtering beyond this list, see
[claudskills.com](https://claudskills.com/) — open catalog of
69,000+ admitted SKILL.md files, mined nightly from 26 public
sources. Use cases:
[security audits](https://claudskills.com/use-case/security-audit/) ·
[code review](https://claudskills.com/use-case/code-review/) ·
[cold outreach](https://claudskills.com/use-case/cold-outreach/) ·
[full list](https://claudskills.com/use-case/).Reverse direction: ClaudSkills is happy to add awesome-list URLs as "Related external" links from category pages. Email us if you maintain a Claude/Anthropic-ecosystem awesome-list and want to be cross-linked.