By tapway
Tapway's collection of Claude Code skills, agents, and hooks for full-stack development with Next.js 14 + Python FastAPI. Includes brainstorming, TDD, code review, systematic debugging, security audit, and more.
Bumps the version in `plugin.json` and `marketplace.json`, adds a new section header to `CHANGELOG.md`, commits, and prints the exact commands to tag and publish the release.
Update all installed Claude Code plugins and their marketplaces to the latest versions. Use this when you want to pull in improvements to skills, agents, or hooks.
Specialized agent for thorough code reviews. Invokes the code-review skill and systematically checks for security, performance, type safety, and compliance with project conventions. Use when reviewing PRs or large diffs.
Specialized agent for infrastructure, CI/CD, deployments, and incident response. Can debug deployment failures, review GitHub Actions pipelines, check Docker configs, and run the deploy checklist. Use for anything ops-related.
Specialized agent for security reviews. Systematically checks code for OWASP Top 10 vulnerabilities, secret exposure, SQL injection, auth bypass, and other security issues. Use before shipping auth, payments, or user data features.
Specialized agent for writing comprehensive tests. Given a file or function, writes unit tests (and integration tests where appropriate) following TDD principles and project conventions. Always writes tests before or alongside implementation.
Fully automated implementation loop: reads a written plan, validates it, executes every task using the TDD subagent pattern (Test Writer → Implementer per task), then runs simplify → review → PR without human intervention. Deploy mode available when running on a staging server: adds deployment, health check, and integration/E2E tests before opening the PR. Standard triggers: "implement it with autopilot", "autoship", "ship this", "execute the plan", "run the plan end to end", "implement end to end", "just ship it". Deploy mode triggers: "implement and deploy", "autoship with deployment", "ship and deploy", "full autoship", "deploy mode".
Explore approaches before writing code for a new feature, architecture decision, or complex bug. Use when there is even a 1% chance multiple approaches exist. Triggers include "let's brainstorm", "how should we approach", "what are the options for", "before we code", or any exploratory design question.
Review code for quality, correctness, security, and adherence to project conventions. Use before opening a PR, when asked to review code, or when self-reviewing changes before committing. Triggers include "review this code", "self-review", "PR review", "check my changes", "before I push".
Safe removal of APIs, endpoints, feature flags, or internal modules with migration support. Requires a tested replacement before any removal, uses Strangler/Adapter/ Feature-flag patterns, distinguishes advisory from compulsory deprecation, and verifies zero active usage before deleting anything. Triggers include "deprecate this", "remove this API", "sunset this feature", "migrate away from X", "clean up old code", "/deprecate".
Adversarial review of high-stakes decisions in-flight — spawns a fresh-context reviewer to disprove rather than validate before a decision is baked in. Applies to service boundaries, irreversible operations, security assumptions, and high blast-radius changes. Triggers include "second opinion", "double-check this decision", "stress-test this design", "am I sure about this", "/doubt".
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
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.
A Claude Code plugin with 17 AI skills, 5 guardrail hooks, and specialized subagents for full-stack development (Next.js 14 + Python FastAPI). The plugin enforces best practices structurally — TDD via agent boundaries, mandatory PR gates, docs on every push — so good habits happen automatically.
# 1. Register this repo as a marketplace
claude plugin marketplace add https://github.com/tapway/tapway-superpowers
# 2. Install the plugin
claude plugin install tapway-superpowers@tapway-superpowers
Skills and hooks activate immediately. No restart, no config files.
Recommended companion plugins:
claude plugin install andrej-karpathy-skills@karpathy-skills
claude plugin install claude-code-setup@claude-plugins-official
Not sure which mode applies? Use this table:
| Situation | Mode |
|---|---|
| Solo developer, building a new feature or fixing a bug | Individual |
| Team of 2+ people working on the same repo in parallel | Team Collaboration |
| Taking over or improving an existing codebase without tests | Legacy Refactor |
Full guide: this section covers it. No separate document needed.
/interview → /brainstorming → /plan → [implement] → /simplify → /review → /pr
/interview is optional but recommended any time the request is underspecified. /observe runs before /pr on any new endpoint. /doubt is called on demand for high-stakes decisions.
Step 0 — Clarify (when the request is fuzzy)
/interview
One question at a time. Stops when it can predict your next three questions. Outputs a Confirmed Intent statement (outcome / user / why now / success / constraints / out of scope) that feeds into /brainstorming or /plan. Never start planning until intent is confirmed.
Step 1 — Brainstorm (optional, but recommended for anything non-trivial)
/brainstorming
Explore approaches, surface trade-offs, name confusion before writing code. Output is committed to docs/brainstorming/. Run /deep-research first if the domain is unfamiliar.
Step 2 — Plan
/plan
Generates docs/plans/[feature].md with file map, task breakdown, and success criteria. For multi-task features, also generates a work package checklist at docs/checklists/[feature]-checklist.md.
Step 3 — Implement
Fast path — fully automated:
"implement it with autopilot" → /autoship
Reads the plan, validates it, runs TDD subagents per task, then simplify → review → PR. You only intervene if a task blocks twice.
Deploy mode — on a staging server:
"implement and deploy" → /autoship (deploy mode)
Same as above, but adds a deployment phase before opening the PR: detects the deploy method (docker compose / make / npm), deploys, waits for a health check, runs integration and E2E tests, then includes the deployment evidence in the PR body. Never opens a PR with a broken deployment.
Manual path — step by step:
/tdd
Test Writer agent (RED) → coordinator gate → Implementer agent (GREEN + REFACTOR) per task. Structurally impossible to skip the failing test.
Between steps — on demand
/doubt
Use this any time you're about to commit to an irreversible decision: service boundary, schema migration, auth flow, security assumption. Spawns a fresh-context adversarial reviewer that sees only the artifact + contract — never your reasoning. Max 3 cycles.
/observe
Run this before /pr on any new endpoint, background task, or external integration. Defines on-call questions first, then adds structured logging (with correlation IDs), RED metrics (bounded labels), OpenTelemetry auto-instrumentation, and symptom-based alerts. Includes staging verification: induce failures, confirm telemetry catches them.
npx claudepluginhub tapway/tapway-superpowers --plugin tapway-superpowersHarness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
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.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Complete developer toolkit for Claude Code