Every "awesome skills" list tells you a skill is good. Few show the source footprint, install tradeoffs, or what real users report after installing it.
New reviews use the skill source, repository metadata, public issue history, commit history, and direct community links. Legacy August 24-25 entries remain marked as measured runs.
Context figures in new reviews are labelled as file-size estimates: bytes divided by four. They are not measured bills and do not come from private traffic.
Dollar figures use published list pricing: Claude Sonnet 5 at $2 per million input tokens and $10 per million output tokens, Claude Opus 5 at $5 and $25, Claude Haiku 4.5 at $1 and $5. Multi-turn sessions re-send history each turn, so input tokens grow with conversation length. That's included in our totals.
Addy Osmani's source-first workflow makes assumptions, boundaries, tests, and human approval explicit before implementation.
101,218 stars · 25 skills · verdict: strong for consequential work, ceremony for tiny fixesThe 149k-star lazy-senior-dev skill, read from source. Its hooks inject ~1,313 estimated tokens into every session and subagent; GitHub issues show where shortest-diff-wins goes wrong.
149,188 stars · 6 skills · verdict: cuts over-building, can cut requested work tooThe one script-free skill in the ui-ux-pro-max repo. We read both files, the issue tracker, and the published npm tarball instead of running it through our API.
131,332 stars · 7 skill folders · verdict: useful HTML-first brief, npm copy still pre-fixAnthropic’s frontend-design skill is 9,390 bytes of taste-first instruction. We read the source and public GitHub issue history instead of running it through our API.
174,864 stars · 19 skill folders · verdict: strong taste, blank-canvas biasA 16,964-byte native OpenClaw office-hours skill from gstack. We read the source and public GitHub issue history instead of running it through our API.
130,481 stars · 119 SKILL.md files · verdict: strong product discipline, YC-coded caveatsA 12,493-byte web-app security checklist from ECC, plus a 10,171-byte cloud security companion. We read the source and public GitHub issue history instead of running it through our API.
243,763 stars · 286 README-listed skills · verdict: useful checklist, real harness caveatsA 9,465-byte root-cause debugging skill from the Superpowers framework. We read the full 40,772-byte directory and public issue history instead of running it through our API.
278,122 stars · 14 skills · verdict: strong discipline with real activation rough edgesAnthropic's own API reference skill ships a 75,707-byte SKILL.md and a 973,337-byte source set. We read the source and the issue tracker instead of running it through our API.
171,582 stars · 20 skills · verdict: authoritative docs with real context-risk complaintsA 3,549-byte skill that enforces red-green-refactor one slice at a time. We drove a five-round session adding a coupon feature and measured every turn against an unguided control arm.
5-round session 28,214 tokens · $0.074 on Sonnet 5The most popular skill in a 234,000-star repo is 1,987 bytes of markdown that makes the agent interrogate you before it writes code. We ran a four-round session on a real feature spec and measured every turn.
4-round session 13,519 tokens · $0.037 on Sonnet 5| Skill | Source | Evidence base | Context / tokens | Signal | Criticism |
|---|---|---|---|---|---|
spec-driven-development | addyosmani/agent-skills | source + GitHub/HN signals | 13,096 B SKILL.md; ~3,274 est. tokens | 101,218 stars; 25 skills | approval gate can be bypassed; host-specific install boundaries |
ponytail | DietrichGebert/ponytail | source + hooks + GitHub/HN signals | 6,637 B SKILL.md; ~1,313 est. tokens injected per session/subagent | 149,188 stars; 6 skills | shortest-diff tiebreak can drop requested work; always-on incl. subagents; cost claim Haiku-only |
banner-design | nextlevelbuilder/ui-ux-pro-max-skill | source + GitHub issues + npm tarball | 7,071 B SKILL.md; ~1,768 est. tokens (~3,016 with reference) | 131,332 stars; 7 skill folders | npm 2.15.0 ships pre-fix file; thin, unsourced platform specs |
frontend-design | anthropics/skills | source + GitHub issue signals | 9,390 B SKILL.md; ~2,348 est. tokens | 174,864 stars; 19 skill folders | blank-canvas bias; weak design-system + handoff path |
gstack-openclaw-office-hours | garrytan/gstack | source + GitHub/HN signals | 16,964 B SKILL.md; ~4,241 est. tokens | 130,481 stars; 119 SKILL.md files | YC closing pitch can feel like brand gravity; full workflow has state/handoff caveats |
security-review | affaan-m/ECC | source + GitHub issue signals | 12,493 B SKILL.md; ~3,123 est. tokens | 243,763 stars; 286 README-listed skills | broad checklist; framework has real security/install-surface caveats |
systematic-debugging | obra/superpowers | source + GitHub/HN signals | 9,465 B SKILL.md; ~2,366 est. tokens | 278,122 stars; 14 skills | rigid process can become noise on mechanical fixes |
claude-api | anthropics/skills | source + GitHub issues | 75,707 B SKILL.md; ~18,927 est. tokens | 171,582 stars; 20 skills | broad trigger can waste context |
tdd | mattpocock/skills | legacy measured run | 28,214 measured tokens | 5-round session | not cheaper than control |
grilling / grill-me | mattpocock/skills | legacy measured run | 13,519 measured tokens | 4-round session | can be overhead on trivial tasks |
More skills land here daily. New reviews use source code, public GitHub signals, and attributed community reports rather than private traffic measurements.
Skills are markdown files that get injected into the agent's context. A big skill is a big prompt, and a skill that runs in rounds re-sends its own instructions on every turn. Two skills can look equally useful in a README and differ by 10x in what they bill you.
That gap is invisible until someone measures it. So we measure it. If you want to check our numbers against your own workload, the API cost calculator and token counter are free and need no signup.