Recently added

Two-stage cosmetic label design and GLB production, preview, validation, and export

Portable Bun-powered vault workflows for durable projects, notes, and agent goals

Automatically copies gitignored files and configures direnv/mise when agent worktree isolation creates a worktree

Updates tmux window title with Claude Code status indicators showing current activity (working, editing files, running commands, etc.)

Snapshot the running tmux/claude/codex agents before a reboot and resume the exact sessions afterward, complementing tmux-resurrect which restores windows but not the agents inside them.

Temporal database patterns for system-time versioned tables

Stop hook that catches ownership-dodging and session-quitting phrases and forces the assistant to keep working

Ends every turn with a one-sentence recap of what the user asked for and the URL they are most likely to want next

Development guidelines for writing stable, maintainable skills, agents, and commands

Core orchestration guidelines for conversation style, code style, testing, tool conventions, and workflow automation

Task management system using markdown checkboxes in .llm/todo.md with slash commands for planning, implementing, and tracking tasks across git worktrees

Utilities for working with justfiles, including doc comment optimization

Java and Maven tools for OpenRewrite import ordering and POM dependency management

Updates iTerm2 window title with Claude Code status indicators showing current activity (working, editing files, running commands, etc.)

Explain command-line investigations as self-contained HTML reports with real commands, output, and reasoning

Reports Claude sessions to Herdr and synchronizes custom tab titles.

Snapshot the running herdr workspaces, tabs, and claude/codex agents before a reboot and rebuild them afterward, resuming each agent's exact session.

GitHub Actions troubleshooting and CI/CD automation

Git workflow automation with smart commits, conflict resolution, rebase management, worktree creation, and worktree cleanup

Block dangerous git commands before they execute

Updates Ghostty window title with Claude Code status indicators showing current activity (working, editing files, running commands, etc.)

Code quality tools for comment cleanup and emoji enhancement

Send session lifecycle events to the claude-code-plans dashboard

Task orchestration using Claude Code's built-in task tools with slash commands for importing plans and executing tasks individually, sequentially, or with a team

Test and build automation tools that loop until all commits pass

Block dangerous bash commands before they execute

Development-workflow skills: design, TDD, adversarial review, shipping, and campaign orchestration.

Draft, schedule, and manage social media posts across X, LinkedIn, Instagram, TikTok, YouTube, Threads, Bluesky, Facebook, and Pinterest

One-line description of what this skill does

Send messages between local Claude Code sessions. Prefers pushing straight into the target session as a channel event; falls back to MCP tool results and end-of-turn hook context where channels are unavailable. Never uses the input box.

Canonical rooms, task routing, context, receipts, and artifacts for Claude, Codex, Gemini, and LLM Caret.

SPipe guarded release planning, process guidance, and MCP tools.

Personas, workspaces and memory for coding agents across many repos — Claude Code, opencode and Codex

Audit the health of all rainlanguage GitHub org repos and produce a prioritized modernization report.

YNAB MCP connector for budgets, transactions, categories, scheduled transactions, accounts, months, payees, and budget review workflows.

YNAB MCP connector for budgets, transactions, categories, scheduled transactions, accounts, months, payees, and budget review workflows.

Curate any Markdown knowledge vault: catalogue, classify, audit, repair frontmatter, rename with link rewrite, check broken links, inject emoji sigils, infer tags, and detect themes. Works with Obsidian, mkdocs, Jekyll, or any .md tree with YAML frontmatter.

Spec-Driven Development for research papers: proposal, related work, feasibility, plan, one task queue executed through implement, analyze, and review.

Interactively generates a production-ready CLAUDE.md and .claude/ structure for a project.

Joe's engineering workflow toolkit: skills and agents for TDD, systematic debugging, planning, code review, and shipping.

원격 fintool MCP와 시장·기업 재무 스킬 6종(valuation·financial-statements·fund-performance·portfolio-risk·fixed-income·startup-finance), 스타트업 검증 패키지 4종(startup-design·competitors·positioning·pitch). 한국 시장 보정 references 포함.

신동형 인사이트 4종 문서 생성기 (Word KR+EN, PowerPoint KR+EN). v0.9.1 Unified Deep-Verify Edition: 동적 워크플로(ultracode) 하니스 내장, Deep Verification(Claims Ledger 기반 할루시네이션 차단), So-What 인사이트 프레임워크, 시각화 카탈로그, Word 챕터 자동 번호(H1C/H2C + HEADNUM 게이트), 4중 품질 게이트(toc/content/claims/judge).

See and drive the Windows desktop the way the Playwright plugin drives a browser — screenshot and control tree come back with every action — and record the session into a finished demo video with click-zoom, a smoothed cursor and narration.

Persistent memory for AI coding agents. Captures your sessions, consolidates them into a searchable markdown wiki, and hands the next session a short list of pointers. One binary, nothing resident.

Analyze incoming-call impact and manage function notes through the Impact Lens Agent CLI.

Self-audit AI agent, tool, and MCP-server code for security and reliability misconfigurations with Trustabl, the agent reliability scanner for the OpenAI Agents SDK, Claude Agent SDK, Google ADK, and MCP. Ships two skills: trustabl-scan scans your agent code with Trustabl (via a bundled MCP server) right after you write or change it, before you commit — optionally matching your declared dependencies against the OSV database for known CVEs — and trustabl-enrich applies the scan findings directly to your source files as targeted code edits.

Evidence-first reading of supplied documents — books, articles, PDFs, document sets. Five modes (quick, deep, map, feynman, book), claim/reason/evidence decomposition with four-level confidence labels, an evidence ledger, knowledge maps, and Feynman teach-back. Starts from material the user supplies; for discovering sources across the web use deep-research, and for citation-formatted briefs use research-summarizer.

CEFR-leveled Swedish-learning mentor. Places the learner with a two-question probe (never trusting a vague self-label), builds a listening/reading/writing/speaking path, and recommends 3-6 vetted YouTube and podcast resources by level — Peter SFI, Lätt Svenska med Oskar, Radio Sweden på lätt svenska, Klartext, Fluent Fiction, Sommar i P1 — with a maintained catalog in references/swedish-resources.md and a hard never-invent-a-URL rule.

Organic LinkedIn presence, end to end, with the platform rules enforced in code. An orchestrator (context: fork) gates every request against LinkedIn's User Agreement §8.2 — refusing automation, scraping, engagement pods, bulk DMs, fake identity, fabricated proof, and named third-party automation tools, each with a compliant substitute — then routes to one of five sub-skills. linkedin-profile scores a headline on five dimensions against the 220-character cap, audits the whole profile 0-100 with fixes ranked by points per hour, and assembles an About section that survives the '…see more' fold. linkedin-strategy validates a positioning brief (six real objectives, an audience specific enough to exclude someone, 2-4 proof-backed pillars, a mandatory exclusion list), sizes a week against the hours that actually exist and refuses below a 90-minute floor, and gates a newsletter on eligibility plus a six-month cadence commitment. linkedin-content lints posts 0-100 across mechanics, hook, integrity, and accessibility — blocking on the 3,000-character cap, engagement bait, and Unicode pseudo-bold that screen readers cannot read — picks the format the material actually supports, and splits long sources into standalone units against a reuse ledger. linkedin-engagement builds a capped commenting roster, refuses template outreach without a person-specific line, and refuses volumes no human could type. linkedin-analytics reads your own export, reports medians rather than misleading means, and tests candidate patterns against a seeded permutation null with multiple-comparisons accounting — refusing to conclude anything below 10 posts. 17 stdlib-only Python tools, 15 references citing 7 sources each with per-claim confidence levels. No LinkedIn credentials, no API calls, nothing auto-sent.

Grade an agent setup from real conversation history. Harvests recent local Claude Code and Codex sessions scoped to one repo, condenses and secret-redacts transcripts (12-pattern redactor, chmod-0600 artifacts, nothing ever uploaded), has the agent judge each transcript against two fixed rubrics (efficiency, code-quality — labels only, preserved verbatim from upstream), then runs a deterministic aggregation gate that derives every number from the label tables, refuses scores for unsampled sessions and suggestions that cite no scored session, and renders one self-contained HTML report (no JS, no CDN, dark-mode, print-to-PDF). Proposed skill edits stay staged as diffs until an explicit per-skill yes, and zero suggestions is a valid, reportable success. Ships cs-skill-doctor agent, /cs:skill-doctor command, 3 stdlib tools, 2 verbatim rubrics, 3 references citing 7 sources each, and the 3 handoff-shape assets.