start-from-here
claudecodexInterviews you, then writes approved pre-code project docs (PRD, TRD, architecture, stack, security, tests, legal, AGENTS.md) before any code is written.
1 skill
Claude Code and Codex plugins their authors filed under "productivity". All categories.
Interviews you, then writes approved pre-code project docs (PRD, TRD, architecture, stack, security, tests, legal, AGENTS.md) before any code is written.
1 skill
Renders Mermaid diagrams (32 types) to SVG or HTML with Pele, without a browser.
1 skill
Local conversation tree with a native MCP App panel.
1 MCP server
Export the complete accessible current conversation and workspace state to a structured Markdown handoff.
1 skill
Investigate open-weights language models in a persistent local worker.
1 skill
Windows host maintenance: disk-saver diagnoses disk usage with usage evidence (running apps, VS Code extension lists, nvm/Playwright references, superseded drivers, pagefile, stale workspaces), proposes prioritized Japanese reports, and removes only per-item approved targets
1 skill
my-swift-linter CLI and rule configuration guide for AI agents and humans.
Tri-CLI orchestrator for Claude Code. Routes tasks across Claude + Codex + Antigravity with consensus modes, live Claude↔Codex sessions, worktree-isolated swarms, and cross-model review.
Compacts when the work moves on from one sub-task to the next instead of when the context window fills, including in the middle of a long turn. Once the context is past a floor, a model judges each step, reading what you asked for and where Claude has got to but never any tool output, for whether the work moves on from a finished piece to another; if it does, taskcut runs Claude Code's own compaction. Below the floor it does nothing and costs nothing.
uwuclxdy's skill collection.
Schedule, manage, and analyze social media posts via PostFast. Supports Facebook, Instagram, X, TikTok, LinkedIn, YouTube, BlueSky, Threads, Pinterest, Telegram, and Google Business Profile.
Read-only side chat: /aside opens a pane where you ask about the session so far, answered by a tool-less fork of the transcript; nothing is written back into the main thread.
사용량 band 표시와 서브에이전트 effort 정책 강제
One pane for the whole session: the original intent and drift, subagents and background shells with model, tokens, idle and typical duration, files changed, and the last reply with its open questions.
Click a .md path in the Claude Code conversation to read it rendered as markdown beside the session, pictures included, and point at any block to have Claude edit it.
Sets the terminal tab title to reflect current Claude Code activity.
Strips wasteful patterns from Bash tool output (git boilerplate, long lines, repeated diagnostics) to save tokens.
Spawns steerable child Claude Code sessions in terminal windows with automatic result forwarding via hooks. Requires the spawn-claude CLI to be on PATH.
Never get surprised by a rate limit again. An always-on band above the prompt shows your 5-hour and weekly Claude usage with live bars, reset countdowns and context-window fill. It turns red when your pace will run out before the reset, and warns you at 75%, 90% and 100%.
Come back to big chats for less. When a large chat (150k+ tokens) has been idle for 25 minutes, it is compacted while the prompt cache is still warm, so picking it up later is cheap. The summary keeps your last request, open tasks, decisions, file paths, commands and errors.
my dotfiles
GitHub Copilot variant — agents emitted as <name>.agent.md with camelCase hook events
170 skills · 70 commands · 29 agents · 1 MCP server · hooks
Universal governance — agents, skills, commands, hooks, and rules for all projects
170 skills · 70 commands · 29 agents · 1 MCP server · hooks
Cursor (cursor-agent) variant — skills and agents consumed via Cursor's native .claude-plugin loader
170 skills · 70 commands · 29 agents
Your Virtual CTO — 60 AI agents. 265 expert skills. 15 quality gates.
102 skills · hooks
Kelly's App-in-Skill business tools: email inbox-zero, money ledger, finance anomaly audit, personal CRM, unified chat inbox, inquiry-to-quote desk, ticket triage & dispatch, lesson-plan compliance, social media aggregation, Search Console SEO desk, user feedback roadmap, market intelligence radar, fleet ops monitoring, content repurposing, GitHub PR review, short-drama & music-video production, team standup board, cross-border e-commerce (product research, listing factory, ads desk), and multi-agent rule setup. Each bundles a local browser UI for human review and approval.
84 skills
Run /kronael:sync (or say 'sync kronael') to deploy language patterns, workflow commands, wisdom skill, and lifecycle hooks into ~/.claude/. Re-run to update.
70 skills · 1 command · 5 agents
Find and score jobs, tailor applications, prep and mock interviews, learn your preferences, track everything on a local dashboard.
34 skills · 22 commands · 6 agents · 1 MCP server
Structured knowledge capture, lesson-learned documentation, and cross-session memory for Claude Code projects. Includes KG entries, ADRs, meta-issue tracking, chat extraction, and profile-file behavioral memory.
16 skills · 31 commands · 9 agents · 1 MCP server · hooks
PARA-style Obsidian vault management with frontmatter validation, template versioning, and git auto-commit. Provides 20+ MCP tools, 15 slash commands, and 7 skills for complete vault operations including AI-assisted content injection and progressive summarization distillation.
23 skills · 20 commands · 1 agent · 1 MCP server · hooks