🔌
Plugins
Loading...
Showing 30 of 23020 results
Launch and track parallel work in isolated git worktrees. Prevents duplicate work, tracks active branches.
平行派發任務給多個 AI agent（Claude + Codex），獨立執行後交叉比對結果
서브에이전트 병렬 위임 워크플로 + 3-role 에이전트(구현/테스트/구현품질검증): 독립 작업을 단일 응답에 다수 Agent 호출로 묶어 동시 실행. 자기 검증의 함정 차단을 위한 역할 분리 패턴 + 카탈로그 자동 발견·description 매칭 메커니즘. 누락 역할만 보충 설치하는 setup 스킬 포함(사용자 정의·OMC 에이전트 우선). 2개 스킬 + 3개 에이전트.
Time tracking and session management powers for agentic development workflows. Track development time, log AI usage, and manage GitHub issues — all from within your coding agent.
Write and review Paramiko code with explicit host-key policy, exit-status checks, and deadlock-safe stream handling — no leaked connections.
English skills: evidence-disciplined testing framework, Playwright test review, test cases, Jira bug reports
Русская версия скиллов: дотошное тестирование с доказательной дисциплиной
Talk to your other Claude Code projects from any session.
Said it once? Done. Cut 30-40% wasted tokens from repeated ideas.
Your desktop AI reads and saves details from your encrypted ParrotIt vault on demand — decrypted only on your device, every read and save logged, revocable any time. Sensitive data and saving are off by default.
Search companies, analyze product catalogs, discover top sellers, and get pricing intelligence across e-commerce markets.
Graduate-grade physics/materials engine for a particle sandbox: governing equations, canonical parameters (Lennard-Jones, Van der Waals, CODATA 2022), integrator selection, an executable conservation-law validation harness, and bundled/fetchable open datasets with a licensing manifest.
The partnercron package for the partner / channel manager, licensed through salesascode.com.
Build PAS-compliant CLI tools that work seamlessly with AI agents. Includes comprehensive guides, templates, and examples for implementing the POSIX Agent Standard.
Enforces the architecture boundary between PatchLang (.patch) and the JSON layout sidecar — use when touching device data, connection types, canvas scene store, emitter, or loader
专利 PDF 批量下载工具。支持多平台（Google Patents 优先），自动处理申请号和公告号格式
Enforces a single rule: every path written into a repo must be relative to the repo root. Ships a SessionStart directive plus pre-commit and commit-msg git hooks that hard-block commits whose staged files, message, or branch name reference anything outside the repo.
Map the codebase. Pick the path. Forge the goal.
Strip the AI packaging, keep the meaning. Detect and rewrite AI writing patterns across KO/EN/ZH/JA with MPS/fidelity checks. Runs the root SKILL.md as the /patina skill.
Claude Code agents and skills for Patrol-based mobile UI test automation. Generate, author, and debug Patrol integration tests for Flutter apps — with wait-strategy, soft-assert, and flakiness-hardening patterns.
Reusable design patterns: gen/eval, executor abstraction, autopilot orchestration, dynamic workflows, workflow-feedback, design-lens (HCI theory + heuristic health-check).
Harness engineering plugin for systematic AI-augmented product development. 8 core agents + 14 commands wrap a conversational SDD/TDD lifecycle with EXIT GATE verification, ontology-typed task graphs, and zero-borrow self-contained skills. v1.7 adds Decision Briefing: at conversational decision points paul-os renders a self-contained HTML report (context + every decision + options/tradeoffs/recommendation) and opens it in the browser, so the user skims everything at once and answers in one pass — reading lives in HTML, answering stays robust (batched AskUserQuestion or paste-back form). Brand-neutral by default with theme injection. v1.6 added Executable Verification (Wave Checkpoint: runtime-smoke + visual gate + path/symbol). /paul-os:migrate upgrades stamped artifacts across version bumps.
Paul Baders persönliches Skill-Kit: Schreiben in Pauls Stimme (voice-Familie), IMP-Identität, Portrait-Prompts, Projekt-Bootstrap, Ernährungs-Tracking und Gesprächs-Vorbereitung (call-prep). Lernende System-Agenten (outreach, scouts, kleya) bleiben bewusst Code-only — nur zustandslose, MCP-/web-native Skills werden geteilt, siehe CLAUDE.md.