Agent Skills

agent-skills

A personal shelf of skills for coding agents. Hand-crafted, one skill at a time. Specialized skills for coding agents, each one narrow, opinionated, and evaluated before it ships. Includes screenwright, which builds a UI from a design spec, then renders it, audits accessibility, and fixes what does not match.

---

TL;DR

• What this is — a single Claude Code plugin that installs a curated, growing shelf of specialized skills in one go.
• Who it's for — anyone on Claude Code or Cowork who wants auto-triggering expertise for a specific job. [Start here](#start-here) to find yours.
• How to start — run the two-line install below. Each skill triggers on its own description when you describe the job — you don't have to memorize them.

Install

Claude Code (recommended) — installs the full plugin: every skill, slash-command entry points, and one-step updates.

/plugin marketplace add sorawit-w/agent-skills
/plugin install agent-skills@sorawit-w

That's it — every skill below is now on the shelf. Works from Claude Code, Cowork, and OpenAI Codex (≥ 0.133.0) — the same /plugin marketplace add + /plugin install commands run on Codex, then /reload-plugins. The repo ships native manifests for both: .claude-plugin/ for Claude Code and .codex-plugin/plugin.json + .agents/plugins/marketplace.json for Codex.

When the plugin updates, refresh once and reinstall:

/plugin marketplace update sorawit-w
/plugin install agent-skills@sorawit-w

> Claude Code caches the marketplace index locally — new skills and fixes only appear after an explicit refresh.

Any agent — via npx skills. Works with Cursor, opencode, Claude Code, and any tool that reads SKILL.md files — no plugin system required.

npx skills add sorawit-w/agent-skills                          # the whole shelf
npx skills add sorawit-w/agent-skills --list                   # browse, don't install
npx skills add sorawit-w/agent-skills --skill team-composer    # just one

This copies the skill folders — SKILL.md plus their references/, templates/, and hooks/ — into your agent's skills directory. It does not carry the Claude Code plugin wiring: slash-command entry points and MCP/hook registration come only via the marketplace path above.

> Most skills are self-contained and install cleanly on their own. A few share resources — define uses i18n's locale data; startup-grill, sub-agent-coordinator, and wear-the-hat use team-composer's role personas. startup-audit goes further: it has hard dependencies on team-composer, validation-canvas, and riskiest-assumption-test and will refuse to run if they're absent — install the full set for it. Install shared skills alongside their sibling (or just grab everything) so the references resolve.

---

Start here — by what you're doing

New here? Find your job and jump in. The complete map is in [docs/skill-graph.md](docs/skill-graph.md); each skill links to its own README for the full usage, design notes, and cross-skill pairings.

Building & validating a startup

Idea to investor-ready — and each step runs standalone, so jump in anywhere.

brand-workshop → validation-canvas → riskiest-assumption-test → pitch-deck → startup-grill

full chain in [How skills chain](#how-skills-chain).

[brand-workshop](skills/brand-workshop/README.md) — Discovery → Concept → Creation: ships a brand identity package (logo, tagline, brief, DESIGN.md tokens, optional generative mascot + character sheet).

Try:

• "Run the full brand workshop for my meditation startup."
• "Design a mascot and character sheet for my brand."

[validation-canvas](skills/validation-canvas/README.md) — interview block-by-block into a rigorous Lean Canvas + Value Proposition Canvas.

Try:

• "Build a validation canvas for my AI code-review tool."

[riskiest-assumption-test](skills/riskiest-assumption-test/README.md) — turn the canvas's stress-tests into falsifiable hypotheses with a risk×impact matrix + test plan.

Try:

• "I just shipped my validation canvas — what should I test first?"

[pitch-deck](skills/pitch-deck/README.md) — structured interview → investor-ready self-contained HTML deck.

Try:

• "Investor wants my seed deck by Friday — start the structured interview."

[startup-grill](skills/startup-grill/README.md) — adversarial panel → kill report with an Investable / Pivot / Pass verdict.

Try:

• "Grill my startup idea: a B2B SaaS for accounting firms — what would kill us?"

[startup-launch-kit](skills/startup-launch-kit/README.md) — orchestrator (opt-in): runs the five steps above end-to-end with shared state.

Try:

• "Build my whole startup kit for [idea] — take me through everything."

[gtm](skills/gtm/README.md) 🚧 BETA — after the pipeline: phased go-to-market playbook, content, and scheduling.

Try:

• "My pitch deck is ready and I want users — set up GTM for this project."

[startup-audit](skills/startup-audit/README.md) — already built? point at a codebase/URL → fast Continue / Pivot / Kill verdict + R/A/G band, grounded in a build-vs-claim diff.

Try:

• "Grill my startup from this repo — continue, pivot, or kill?"

Reviewing a product or AI feature

Multi-perspective critique and design-completeness checks.

[team-composer](skills/team-composer/README.md) — assemble a virtual team and run a 3-round discussion that forces real disagreement.

Try:

• "Bring a team together to review this mobile auth architecture before we ship."
• "Brainstorm a habit-tracker for teens — multi-perspective, no single viewpoint wins."

[storm](skills/storm/README.md) — perspective-driven, retrieval-grounded research → a cited briefing; the grounding hand-off target for team-composer. (Single-source-of-truth fact-checking → deep-research.)

Try:

• "Research what's actually known about four-day work weeks — grounded briefing, multiple angles."
• "Survey the evidence on RAG vs. long-context for production agents, with sources."

[ai-ux-review](skills/ai-ux-review/README.md) — human-AI design-completeness review; seven blocks, explicit gap markers.

Try:

• "Review the UX of our LLM email-draft feature before launch — focus on trust."
• "We shipped an agentic feature and fluent-but-wrong outputs are surfacing — walk us through the review."

[ai-eval-review](skills/ai-eval-review/README.md) — the eval-design sibling of ai-ux-review, plus a regulatory lens.

Try:

• "Review the eval setup for our classification model — push on ground truth and cohort breakdown."
• "Eval review for our medical-imaging classifier — we're FDA SaMD class II, apply regulatory rigor."

[gamification-fit](skills/gamification-fit/README.md) — restraint-first: finds the few places play honestly fits, refuses the rest.

Try:

• "Where should I add gamification to my habit-tracker so users log 4+ days a week?"
• "Should we gamify our expense-report flow? Goal is 90% on-time."

See also: [validation-canvas](skills/validation-canvas/README.md) and [startup-audit](skills/startup-audit/README.md) (under Startup) both double as review tools.

Writing & shipping code

Discipline, stack choices, and parallel work.

kerby (external) — opinionated agentic-coding guardrails (clarity over cleverness, safety over speed, nothing unproven passes the gate). Formerly coding-rules in this repo; now its own repo.

[tech-stack-recommendations](skills/tech-stack-recommendations/README.md) — opinionated default TS/JS stack, with named alternates.

Try:

• "Starting a SaaS side-project — give me one opinionated stack I don't have to second-guess."
• "We're migrating off Next.js on Vercel — recommend the path and name the trade-offs."

[sub-agent-coordinator](skills/sub-agent-coordinator/README.md) — orchestrate parallel sub-agents without drift.

Try:

• "Refactor all 14 React components from class to function — coordinate in parallel."
• "Debug our flaky CI: spawn a researcher, a fixer, and a reviewer with clear briefs."

[wear-the-hat](skills/wear-the-hat/README.md) — do a task in one expert's voice, solo, no panel.

Try:

• "Audit middleware/auth.ts for missing CSRF guards as @security_specialist."
• "Wear the accessibility specialist's hat for this UI review."

[screenwright](skills/screenwright/README.md) — paints one self-contained HTML surface, then renders it via the Playwright MCP and fixes it against an axe + fidelity gate until it passes (needs the Playwright MCP).

Try:

• "Build this dashboard card to match the mockup and verify it's accessible."
• "This component looks off on mobile — screenwright it and check the render."

Calibration & personal voice

Make the agent yours, and write as yourself.

[whoami](skills/whoami/README.md) — profile how you want the agent to collaborate with you (six dials, an RPG class, a portable sheet)

Try:

• "Run /whoami — I want you to actually know how I like to work."
• "/whoami rerun — my role changed and the profile's out of date."

[handshake](skills/handshake/README.md) — a brief calibration ritual before real work begins.

Try:

• "Calibrate how we work — I'm tired of generic answers."
• "Run /handshake --project for this repo before the next sprint."

[ghostwriter](skills/ghostwriter/README.md) — draft messages in your own voice, zero AI tells.

Try:

• "Reply to this email from my manager — keep it short."
• "Tell my teammate on Slack the deploy slipped to Thursday, style=friend."

[plain-english](skills/plain-english/README.md) — install a plain-English rule set once, and every later reply stays readable for a second-language reader: common words, the answer first, no idioms, technical terms untouched.

Try:

• "Your replies are too long and too hard to read — install the plain-English rules."
• "Rewrite this paragraph so a non-native English reader can follow it."

Language & visuals

[i18n](skills/i18n/README.md) — surgical edits on large translation files, plus cultural rewriting.

Try:

• "Translate this onboarding flow into Thai, Japanese, and Korean — cultural rewriting, not machine translation."
• "Edit three keys in a 4,000-line zh-CN.json without rewriting the whole file."

[define](skills/define/README.md) — the true in-context meaning of a word or phrase, with a learner gloss.

Try:

• "What does ออเจ้า mean in 'ออเจ้าจักไปไหน'? Give me the register and why."

[pixel-art](skills/pixel-art/README.md) — a pocket pixel-art studio with a built-in design system: generates a single subject, quantizes it to a true grid, returns PNG or single-sprite SVG (or a model-agnostic prompt brief). Repo-agnostic; declines composed/branded asset jobs instead of faking them.

Try:

• "Create a hi-fi pixel-art medieval harbor at dusk with a lighthouse and three ships."
• "Generate a pixel-art knight sprite and give it to me as SVG."

Authoring & auditing skills

Building skills for this shelf? [skill-evaluator](skills/skill-evaluator/README.md) audits whether a SKILL.md's rules actually land when Claude runs it — pair it with [team-composer](skills/team-composer/README.md) to design and [sub-agent-coordinator](skills/sub-agent-coordinator/README.md) to build. Full authoring guide in [Building on the shelf](#building-on-the-shelf).

Each skill lives under [skills/](skills/) with its own README.md, SKILL.md, and reference docs.

This repo treats skill authoring as harness engineering — designing the context, scaffolding, feedback loops, state, and eval discipline around an agent so it can do reliable work. See [CLAUDE.md](CLAUDE.md) → "Harness vocabulary" for the five primitives and where each shows up in the shelf.

---

How skills chain

Two pipelines the shelf is designed to support end-to-end.

🧭 Startup pipeline — identity → beliefs → tests → deck → grill → ship 🚧

brand-workshop ──▶ validation-canvas ──▶ riskiest-assumption-test ──▶ pitch-deck ──▶ startup-grill ──▶ gtm 🚧
 (identity kit)    (Lean Canvas + VPC)    (test plan + results)        (HTML deck)    (kill report)    (BETA — get users)

The sixth step (gtm) is beta — it slots in after startup-grill for founders who've graduated the pipeline and want to actually go acquire users. Evals validate structural reliability; real-world dogfooding is the next milestone before v1.

Sequential by default. No one-shot orchestrator — each skill is invocable independently, but composes through filesystem conventions. Inter-step gates are weighted: brand-workshop → validation-canvas is light; validation-canvas → riskiest-assumption-test is medium (RAT STOPs without canvas); riskiest-assumption-test → pitch-deck is heavy (pitch-deck STOPs without populated ## Results for top-3 hypotheses; override with [PRE-VALIDATION DRAFT] watermark); pitch-deck → startup-grill is light. Loop-back is first-class — invalidated hypotheses route back to validation-canvas in update mode, not a pipeline restart. Pristine pipelines (no canvas revision after testing) are the actual yellow flag, which startup-grill checks for in its kill-report ## Iteration Evidence section.

The artifacts compound. As of v2.2.0, every artifact lives under docs/ (default docs// solo, docs/startup-kit// orchestrated; v1 layouts at cwd root still read via fallback). brand-workshop writes docs/brand/DESIGN.md (Google Labs spec, alpha) — validation-canvas, riskiest-assumption-test, and pitch-deck all pick it up automatically and extract tokens from the YAML front matter for consistent branding. validation-canvas writes docs/canvas/validation-canvas.md (Lean Canvas + VPC) — its Stress Tests section seeds riskiest-assumption-test's assumption dump; pitch-deck reads it to seed slides 2, 3, 6 and cross-checks the Ask against the Stress Tests. riskiest-assumption-test writes docs/rat/assumption-test-plan.md + interactive docs/rat/test-matrix.html — pitch-deck reads ## Top 3 Hypotheses and ## Results to inform the Validation slide and Traction claims; startup-grill reads them for the iteration-evidence check. startup-grill reads everything as direct grilling ammunition and ships a docs/grill/kill-report.md with a verdict you can act on. You don't have to wire anything up; running them in order is the wiring.

Pipeline philosophy: validation is iterative, not a checklist. Each step's value is in the slow consideration it forces.

Pipeline shortcut (v2.1.0+): the optional startup-launch-kit skill is an opt-in orchestrator that sequences the five steps with shared state via kit-manifest.json (at docs/startup-kit/kit-manifest.json as of v2.2.0). It is convenience, not replacement: every individual skill remains independently invocable, every gate is honored (no silent bypass), and every step still surfaces its own prompts to the founder (no batching). The orchestrator preserves the philosophy by recording every gate override with a reason, by surfacing loop-back recommendations after RAT and grill (founder decides; never auto-routes), and by treating the manifest as a hint rather than truth (filesystem state always wins on reconciliation). Use it when you want end-to-end coordination without manual step-chaining; use the individual skills directly for any single-step or partial-pipeline work.

🛰 Delegation pipeline — discuss → build

team-composer ──▶ sub-agent-coordinator
 (3-round debate)   (parallel build with verification)

team-composer's Phase 6 hands its conclusion, decisions, and role constraints to `sub

…

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: sorawit-w
• Source: sorawit-w/agent-skills
• License: MIT

Install and usage instructions live in the source repository linked above.