By dotts-h
Methodology skills for software projects: recording-decisions (ADRs), practicing-tdd, registering-contracts, maintaining-conventions, logging-learnings, managing-tech-debt, mapping-codebases, improving-architecture, auditing-code-quality, hardening-tests, authoring-tests, exploring-quality, designing-ui-ux.
Reviews code against a patterns/antipatterns catalog — Go idioms plus this project's pure-core/thin-edges and seam discipline, error handling, naming, and simplicity — and proposes focused cleanups. Use this whenever the user asks to review code quality, enforce coding standards, spot antipatterns, or clean up a diff for clarity. It complements the built-in /code-review (bug-finding) and /simplify (mechanical cleanup) rather than duplicating them.
Writes and enhances the higher test layers — end-to-end (Playwright), API/contract, performance/benchmark, and accessibility tests. Use this whenever a feature needs browser, integration, api, perf, or a11y coverage, when the user asks to add or extend e2e/api/perf tests, or after exploratory QA surfaces a behavior worth locking in. For unit/seam tests written test-first, use practicing-tdd instead.
Audits and improves both UX (Nielsen/Krug heuristics, information architecture, user flows, WCAG accessibility) and UI (visual hierarchy, design tokens, typography, spacing, color, motion — Refactoring-UI principles), then implements the changes in the app's HTML/CSS and verifies them in a real browser. Use this whenever the user wants the interface to look or feel better, more modern, more usable, or more accessible — "make the UI nicer", "the UX feels off", "improve the design", "fix accessibility" — even if they don't name a specific screen.
Runs exploratory QA in two phases — phase 1 generates and runs many breadth probing scripts derived from the codebase (route/HTTP sweeps, smoke flows, malformed-input fuzzing) headless; phase 2 deep-dives the live app in a real browser via Playwright (MCP/CLI), following curiosity. Produces a ranked findings report. Use this whenever the user wants exploratory testing, a bug hunt, a pre-release sweep, or to "poke at the app and see what breaks".
Audits and hardens an existing test suite (the SDET role) — finds coverage gaps, weak assertions, flaky and non-deterministic tests, and missing edge/property/fuzz cases, then strengthens them. Use this whenever the user asks to review, harden, validate, or improve test quality, hunt flakes, check coverage, or before trusting a suite to gate a release. Do not write product code here — only strengthen tests.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
ori is my personal Claude Code plugin marketplace (served from this repo), so my
tooling is available everywhere — including Claude Code on the web/phone (cloud
sessions install declared plugins at session start). It hosts two plugins:
| Plugin | Install id | What it is |
|---|---|---|
skills | skills@ori | 13 reusable methodology skills (docs, process, QA, design). |
cookbook | cookbook@ori | Recipe system: install versioned doc-based engineering-process packages into any repo, update via a lock + 3-way merge, check conformance with doctors. |
Add to the repo's committed .claude/settings.json:
{
"extraKnownMarketplaces": {
"ori": { "source": { "source": "github", "repo": "dotts-h/claude-skills" } }
},
"enabledPlugins": {
"skills@ori": true,
"cookbook@ori": true
}
}
enabledPluginsis an object keyed byplugin-id@marketplace-idwith boolean values — the array form (["skills@ori", ...]) is silently ignored by the settings schema, so the plugins never enable.
Commit + push. Any cloud session for that repo installs the plugins automatically at session start (GitHub is in the default network allowlist), provided this marketplace repo is public — the cloud container clones the marketplace source directly from github.com without credentials. Locally you can also add them interactively:
/plugin marketplace add dotts-h/claude-skills
/plugin install skills@ori
/plugin install cookbook@ori
skills plugin13 methodology skills, bundled:
| Group | Skills |
|---|---|
| Docs | recording-decisions · registering-contracts · maintaining-conventions · logging-learnings · mapping-codebases |
| Process | practicing-tdd · managing-tech-debt · improving-architecture · auditing-code-quality |
| QA | hardening-tests · authoring-tests · exploring-quality |
| Design | designing-ui-ux |
cookbook pluginInstallable, versioned, doc-based engineering process packages. A recipe turns a hard-won process (a constitution, quality gates, an issue store, a dev loop, a release ritual, a contracts registry) into something you install into a repo, keep updated, and conformance-check — the way a package manager treats code.
| thing | lives where | lifetime | who benefits |
|---|---|---|---|
| plugin / skill | the agent host | per-session context; leaves no trace in the repo | the agent that loaded it |
| recipe | the target repo (docs, scripts, CI, .recipes/lock.json) | durable state, versioned + updatable | any agent (or human) that opens the repo, on any host |
Skills here are the operators; recipes are the payload. A repo that adopted
recipes works with no plugin installed at all — the docs and scripts are just there,
host-agnostic (both CLAUDE.md and AGENTS.md thin pointers).
# In the repo you want to bring under management (with cookbook@ori enabled)
/adopt-recipes
adopt-recipes inventories the repo, proposes a profile + tier + answers, confirms
with you, installs in dependency order (harvesting existing docs on brownfield —
never clobbering), writes .recipes/lock.json, and reports gaps. Later:
/update-recipes (semantic 3-way merge to new versions, via PR) and /recipe-doctor
(aggregate conformance report).
Without the plugin, everything is runnable by hand:
make evals # run all recipe evals (fixture repos in mktemp)
make lint # bash -n all scripts + secret/date scans
make doctor TARGET=/path/to/repo # aggregate doctors against a repo
plugins/cookbook/scripts/install-recipe.sh \
--recipe plugins/cookbook/recipes/core --target /path/to/repo --tier M
| recipe | layer | what it installs |
|---|---|---|
core | 0 | CLAUDE.md + AGENTS.md thin pointers, CONVENTIONS constitution (doctrine), CONTEXT glossary; ADR log (M+); CODEMAP/RETROS/TECH_DEBT (L) |
quality | 1 | single-run CI (make/npm flavors) with docs-only paths-ignore skip, the self-enforcing workflow guard, coverage-floor parameter, opt-in fuzz-smoke workflow, REGRESSIONS register (M+) |
loop | 1 | the dev loop playbook + start-fresh.sh (verified base) + next-issue.sh (dependency-aware picker) |
issues | 1 | hybrid issue store: markdown source of truth + GitHub mirror; INDEX + frontmatter format; epics via group: |
release | 1 | tag-driven release workflow with verified version resolution; SemVer + verify-after playbook |
contracts | 1 | CONTRACTS registry with Provides/Consumes; fleet constellation.yaml + fleet-doctor.sh cross-check |
npx claudepluginhub dotts-h/claude-skills --plugin skillsInstallable, versioned, doc-based engineering process packages (recipes) with adopt/update/doctor skills — durable repo state any agent can work with, not per-session context.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review