mizchi/skills

65 resources in this repository

🎯65Skills

300

🎯Skills65

A Claude Code skill that provides a methodology for iteratively improving agent-facing instructions by dispatching bias-free executor subagents to run them, evaluating results two-sidedly (executor self-report plus instruction-side metrics), and iterating until improvements plateau.

empirical-prompt-tuning

🎯pi-coding-agent🎯Skill

A skill for embedding the pi-coding-agent SDK into Node.js scripts to spawn agent sessions, stream output, scope tools, and register custom tools without forking the CLI. Covers session setup, TypeBox schemas, plugin packaging, and the choice between SDK and JSON-RPC modes.

pi-coding-agent

🎯playwright-cli🎯Skill

Covers terminal-based Playwright automation including running tests, generating test code via codegen, taking screenshots and PDFs, and CI sharding configuration.

playwright-cli

🎯playwright-test🎯Skill

A skill providing best practices and reference patterns for Playwright end-to-end testing, including test configuration, CI setup with GitHub Actions sharding, and techniques for handling waits, network triggers, and drag-and-drop.

playwright-test

🎯retrospective-codify🎯Skill

A meta-skill that extracts lessons from trial-and-error debugging sessions and codifies them as reusable ast-grep rules, skills, or CLAUDE.md rules. Invoked when the user explicitly asks to turn a hard-won insight into a permanent, reproducible safeguard for future tasks.

retrospective-codify

🎯tech-article-reproducibility🎯Skill

A skill that evaluates the reproducibility of technical articles by simulating a first-time reader attempting to reproduce the work on their own machine. It identifies missing information and serves as a final quality check before publishing tutorials, tool introductions, and hands-on articles.

tech-article-reproducibility

🎯skill-selector🎯Skill

A collection of agent skills for frontend consulting, covering code review triage, CI analysis, code hygiene, dependency health, testing posture, security review, state management, and performance auditing. Distributed via APM (Agent Package Manager) following the agentskills.io open standard.

skill-selector

🎯chezmoi-management🎯Skill

A personal meta-skill for managing chezmoi dotfiles, covering source directory operations, diff/apply workflows, skill addition, and pre-push hooks with pkfire and secretlint integration.

chezmoi-management

🎯conventional-changelog🎯Skill

A reference skill for Conventional Commits and automatic CHANGELOG generation, covering commit format, semver tagging, and tool comparison across release-please, changesets, git-cliff, and towncrier. Supports npm, cargo, Python, Rust, and Go workflows.

conventional-changelog

🎯review-image🎯Skill

A skill for reviewing screenshots and images using OpenRouter vision models via bundled Deno scripts, supporting freeform feedback and strict pass/fail CI gates for visual regression testing.

review-image

🎯cloudflare-deploy🎯Skill

A skill for deploying applications and infrastructure to Cloudflare, with decision trees for choosing between Workers, Pages, and other platform services, plus built-in authentication and configuration guidance.

cloudflare-deploy

🎯node-sqlite-vec🎯Skill

A skill that provides recipes and pitfall fixes for using Node 24's built-in node:sqlite module with the sqlite-vec extension for vector and RAG storage in TypeScript, without relying on better-sqlite3.

node-sqlite-vec

🎯gleam-practice🎯Skill

A skill for writing and reviewing Gleam code on the Erlang target, covering Wisp + Mist HTTP services, OTP-based processes, justfile workflows, gleeunit testing, and GitHub Actions CI.

gleam-practice

🎯ast-grep-practice🎯Skill

A skill for using ast-grep as a project lint tool, covering configuration, fix/rewrite rules, constraints, transforms, testing, and CI integration for structural code patterns that general-purpose linters cannot express.

ast-grep-practice

🎯dotenvx🎯Skill

A skill for managing environment variables with the dotenvx tool, covering .env file encryption/decryption, multi-environment configuration, key rotation procedures, and CI/CD integration with GitHub Actions.

dotenvx

🎯translate-programming-language🎯Skill

A skill for planning and executing language-to-language server or application migrations with behavior parity, providing structured workflows for oracle generation, compatibility layers, shadow testing, and cutover planning.

translate-programming-language

🎯moonbit-practice🎯Skill

A best practices skill for AI code generation in MoonBit, covering proper use of moon ide commands for code navigation, common syntax pitfalls, and correct patterns for variables, functions, error propagation, and async.

moonbit-practice

🎯nix-setup🎯Skill

An agent skill for setting up reproducible development environments via devbox or pure Nix flakes, with preloaded templates for MoonBit, Rust, TypeScript, Python, Haskell, and OCaml, covering direnv integration and GitHub Actions CI.

nix-setup

🎯apm-usage🎯Skill

A reference skill for APM (Agent Package Manager), covering apm.yml manifest syntax, install/uninstall/update commands, lockfile workflows, and dependency management for AI agent skills and MCP servers.

apm-usage

🎯justfile🎯Skill

A reference skill for the just command runner, covering justfile syntax, variables, conditionals, built-in functions, and GitHub Actions integration. Designed for maintaining existing repos that use justfiles.

justfile

🎯upstream-fix-and-pin🎯Skill

Guides the workflow for fixing bugs in upstream dependencies by creating a PR, pinning the downstream project to a git ref SHA, and managing the transition until the fix is merged and published.

upstream-fix-and-pin

🎯gh-fix-ci🎯Skill

A skill that debugs and fixes failing GitHub PR checks by inspecting GitHub Actions logs via the gh CLI, drafting a fix plan, and implementing changes only after explicit user approval.

gh-fix-ci

🎯mizchi-blog-style🎯Skill

A Japanese-language style guide and AI-tone detection rubric for mizchi-authored blog posts on Zenn and dev.to. Uses a subagent evaluation workflow to score drafts on writing style authenticity versus AI-generated tone, with specific patterns to avoid and replicate.

mizchi-blog-style

🎯moonbit-js-binding🎯Skill

A skill for writing MoonBit bindings to JavaScript using `extern "js"` declarations, covering FFI type mapping, JS object wrapping with opaque types, Promise bridging with async functions, moon.pkg export configuration, and null/undefined handling at the JS boundary.

moonbit-js-binding

🎯skill-finder🎯Skill

A collection of agent skills distributed via APM (Agent Package Manager), featuring frontend consulting tools for code review, CI analysis, dependency health, security audit, and performance profiling, along with Node.js and other development utilities.

skill-finder

🎯waxa-eval🎯Skill

A collection of agent skills by @mizchi distributed via APM (Agent Package Manager), including frontend consulting tools for code review triage, CI analysis, dependency health checks, testing strategy, and performance auditing.

waxa-eval

🎯extract-glossary🎯Skill

Extracts domain-specific terms, repository implementation maps, and onboarding Mermaid diagrams from one or more repositories or GitHub organizations.

extract-glossary

🎯aws-vault-mfa-iam🎯Skill

A skill from mizchi's agent skills collection distributed via APM (Agent Package Manager), which includes frontend consulting tools for code review, CI analysis, dependency health, security, state management, and performance profiling, along with Node.js and infrastructure skills.

aws-vault-mfa-iam

🎯sql-schema-audit🎯Skill

Skill

sql-schema-audit

🎯flaker-storage-cache-on-ci🎯Skill

Part of mizchi's agent skills collection distributed via APM (Agent Package Manager), offering frontend consulting skills including review triage, CI analysis, code hygiene, dependency health checks, and testing assessments following the agentskills.io open standard.

flaker-storage-cache-on-ci

🎯aws-ecs-service-connect-ipv6🎯Skill

A skill for diagnosing and working around ECS Service Connect DNS aliases returning IPv6 addresses to IPv4-only Fargate tasks, part of mizchi's agent skill collection distributed via APM (Agent Package Manager).

aws-ecs-service-connect-ipv6

🎯optimizing-descriptions🎯Skill

A collection of agent skills by @mizchi distributed via APM (Agent Package Manager), focusing on frontend consulting workflows including review triage, CI timing analysis, code hygiene, dependency health, testing posture, security review, and state management architecture assessment.

optimizing-descriptions

🎯sql-security🎯Skill

Skill

sql-security

🎯k8s-crd-from-typed-schema🎯Skill

Skill

k8s-crd-from-typed-schema

🎯sql-lint🎯Skill

Skill

sql-lint

🎯cloudflare-access-app-setup🎯Skill

Skill