claude-plugins

Browser automation with a persistent session, driven by the Playwright Python API. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website", "log into", or any browser interaction request.

Whole-repo audit for over-engineering. Like ponytail-review, but scans the entire codebase instead of a diff: a ranked list of what to delete, simplify, or replace with stdlib/native equivalents. Use when the user says "audit this codebase", "audit for over-engineering", "what can I delete from this repo", "find bloat", "ponytail-audit", or "/core:ponytail-audit". One-shot report, does not apply fixes.

Harvest every `ponytail:` comment in the codebase into a debt ledger, so the deliberate shortcuts and deferrals ponytail leaves behind get tracked instead of rotting into "later means never". Use when the user says "ponytail debt", "/core:ponytail-debt", "what did ponytail defer", "list the shortcuts", "ponytail ledger", or "what did we mark to do later". One-shot report, changes nothing.

Code review focused exclusively on over-engineering. Finds what to delete: reinvented standard library, unneeded dependencies, speculative abstractions, dead flexibility. One line per finding: location, what to cut, what replaces it. Use when the user says "review for over-engineering", "what can we delete", "is this over-engineered", "simplify review", or invokes /core:ponytail-review. Complements correctness-focused review, this one only hunts complexity.

Forces the laziest solution that actually works, simplest, shortest, most minimal. Channels a senior dev who has seen everything: question whether the task needs to exist at all (YAGNI), reach for the standard library before custom code, native platform features before dependencies, one line before fifty. Supports intensity levels: lite, full (default), ultra. Use whenever the user says "ponytail", "be lazy", "lazy mode", "simplest solution", "minimal solution", "yagni", "do less", or "shortest path", and whenever they complain about over-engineering, bloat, boilerplate, or unnecessary dependencies.

Ingests a new source (web clipping, AI chat export, raw note, transcript) into an Obsidian-based LLM knowledge wiki. Files the source under _raw/, distills it into one or more _wiki/ pages, updates index and log, and pulls bidirectional cross-references. Triggers when the user asks to "ingest this source", "digest this clipping into the wiki", or runs /wiki:ingest while working in a configured wiki vault.

Runs a structural and content health check on an Obsidian-based LLM knowledge wiki. Detects frontmatter issues, broken or asymmetric cross-references, contradictions, stale claims, tag discipline violations, migration backlog overflow, missing hub pages, and orphan content. Produces a structured lint report. Triggers when the user runs /wiki:lint, asks for a "wiki health check", or asks to lint the knowledge wiki.

Runs non-interactive maintenance on an Obsidian-based LLM knowledge wiki — deterministic validation, semantic health checks, trivial auto-fixes, and a written report. Built to be scheduled (cron/launchd). Triggers when the user runs /wiki:maintain, asks to "run wiki maintenance", or "auto-fix the wiki".

Migrates a legacy file from an Obsidian-based knowledge vault into the LLM-Wiki schema. Analyzes the file, proposes a split into raw and wiki content, writes target files with proper frontmatter, pulls cross-references, updates index and log, then deletes the original. Cluster mode handles thematic groups. Triggers when the user runs /wiki:migrate-file, asks to "migrate this legacy note", or wants to bulk-migrate a thematic cluster into the wiki.

Argues with your notes in an Obsidian-based LLM knowledge wiki — challenges assumptions (devil's advocate), surfaces non-obvious connections, or asks Socratic questions. Optionally saves the result as a wiki page. Triggers when the user runs /wiki:think, asks to "challenge this note", "poke holes in my thinking", "find connections between my notes", or "ask me Socratic questions about X".

Lints and formats all Python files in the project using Ruff. Activates when the user asks to "lint my Python", "fix Python code style", "run ruff", "format my Python files", "check Python quality", or "fix linting errors". Installs Ruff if not present, auto-fixes resolvable issues, and reports remaining violations.

Converts PDF files to structured Markdown with automatic mode selection. Supports simple text extraction (fast) and complex layouts with code/tables (vision). Triggers when the user wants to convert a PDF document, extract PDF content, or transform PDF to Markdown format.

Explains code structure, logic, and design decisions for educational purposes. Acts as a patient programming instructor who adapts to the student's level. Activate when the user asks "what does this code do", "explain this function", "how does this work", "walk me through this", "I don't understand this code", or any question about understanding an existing piece of code.

Delegates codebase or text analysis to Google Gemini 2.5 Pro via the MCP bridge when the task exceeds Claude's practical context limit (~150K tokens) or when a second-model perspective adds value. Activate when the user asks to "analyze this codebase", "analyze with Gemini", "this is too large for Claude", "deep-dive the entire src/ directory", or requests analysis of a PDF or large document collection.

Uses Gemini 2.5 Pro as an independent technical reviewer to compare two approaches and validate design decisions. Implements the "propose with Claude, validate with Gemini" pattern. Activate when the user asks to "compare these two approaches", "get a second opinion on this design", "validate this architectural decision", "should we use X or Y", or "review these two options independently".

Transforms visual inputs (screenshots, UI mockups, architecture diagrams, whiteboard photos, PDFs) into code, documentation, or structured data using Gemini 2.5 Pro's multimodal capabilities. Activate when the user shares an image file and asks to "convert this mockup to code", "generate a component from this screenshot", "what does this diagram show", "extract this architecture", or "turn this whiteboard into code".

This skill should be used when the user asks to "review code with hunk", "start a hunk review session", "add comments to a hunk diff", "annotate a diff with hunk", "use hunk for code review", or when the user mentions the `hunkdiff` CLI tool. Provides structured guidance for interacting with live Hunk diff review sessions via the daemon CLI.

Analyses and resolves Git merge conflicts intelligently using semantic code merging, automated test validation, and root-cause analysis of each conflict. Activate when the user says "I have merge conflicts", "help me resolve these conflicts", "there are conflicts after merging", "git merge failed", or "conflicts between my branch and develop". Supports --dry-run, --no-tests, and --strategy options.

Surfaces new product, feature, and business ideas by cross-referencing the user's knowledge wiki, GitHub repositories, customer or project notes, and current market trends. Writes a structured ideas document to the Obsidian vault. Activate when the user runs "/ideas", asks "what could I build next", "give me new project ideas", "what opportunities am I missing", "brainstorm project ideas", or "what should I work on strategically".

Provides a quick status snapshot of all GitHub repositories in a configured organisation — active, dormant, stalled — plus open pull requests and issues. Terminal output only, optimised for speed. Activate when the user runs "/project-pulse", asks "what is the status of my GitHub projects", "show me my repos", "which projects are active", "project overview", "GitHub pulse", or "what is happening in [org]". Works from any directory.

Produces a prioritised daily briefing by aggregating tasks from Obsidian TaskNotes, Google Calendar, Gmail, Linear, and GitHub, then writes the result to today's daily note in the Obsidian vault. Activate when the user runs "/today", asks "what's on my plate today", "give me my daily briefing", "what do I need to do today", or "morning briefing". Works from any directory without requiring Obsidian to be running.

Collects completed tasks, GitHub activity, calendar meetings, wiki learnings, and Linear progress for the past week, then writes a structured retrospective to the Obsidian vault. Activate when the user runs "/weekly-review", asks "what did I accomplish this week", "weekly retrospective", "summarise my week", or "prepare weekly review". Works from any directory.

Answers a question against an Obsidian-based LLM knowledge wiki. Performs index-first search, reads relevant pages, synthesizes an answer with citations, and optionally files the synthesis back as a new wiki page. Triggers when the user runs /wiki:query or asks a question that should be answered from the curated wiki rather than from raw sources.

Creates comprehensive handoff documentation before a /compact operation, enabling a new agent with fresh context to seamlessly continue the work. Activate when the user says "prepare a handoff", "document before compact", "context is getting too large", "I need to hand this off", "create a handoff document", "end of session notes", or "document the current state for tomorrow". Supports --output and --linear-issue options.

Synchronizes documentation between CLAUDE.md, README.md, and docs/. Resolves documentation inconsistencies across project files with configurable sync rules and whitespace-tolerant comparison.

Create lecture presentations for FFHS in reveal.js (default) or PowerPoint format. Use when user requests presentations for PVA sessions, lecture slides, or teaching materials. Supports both web-based reveal.js and traditional PPTX output.

Creates professional git commits with automated pre-commit checks for Java, Python, React, and documentation projects. Generates emoji conventional commit messages and analyzes staging status. Produces atomic commits following best practices.

Automates pull request creation with branch management, code formatting, and integration with professional-commit-workflow. Supports GitHub CLI, automated PR descriptions, and project-specific formatters (Biome, Black, Prettier).

Background knowledge for using the Gemini Bridge MCP tools effectively. Teaches Claude when and how to delegate to Gemini in a model-agnostic multi-agent setup. Claude-invocable only.

Initializes open-source projects with GitHub best practices and git branching strategy. Triggers when setting up new repositories with standardized configuration, Gradle Kotlin DSL, or Python/uv support.

Manages tasks in Obsidian via the TaskNotes Plugin HTTP API. Provides task creation, status updates, filtering by project or status, and vault-wide scanning. Triggers when the user asks to "show my tasks", "create a task", "what should I work on", "mark as done", or wants to filter tasks by status/project.

Removes indicators of AI-generated text. Use this skill when editing or reviewing text to make it sound more natural and human. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and corrects patterns such as: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em-dash overuse, rule of three, AI vocabulary, negative parallelisms, and excessive subjunctive phrases.

Cleans up git assets (worktrees, branches) and updates task status after PR merge. Supports both filesystem (STATUS.md) and Linear issue tracking. Safe defaults with dry-run preview mode.

Converts text with visual formatting into proper Markdown syntax, fixes Markdown formatting issues, and ensures consistent document structure. Use this skill when formatting or reviewing Markdown files, fixing heading hierarchies, converting visual cues to proper syntax, applying Swiss German orthography to German-language documents, or checking Markdown against linting rules.